Reading their modified license terms, it cracks me up, because they've basically remade the MIT to be the MIT + the one clause that the BSD used to have, which didn't care about MAU or revenue, if you used it in a product, they asked you to 'advertise' them basically. Honestly, its a reasonable request.
htrp 1 days ago [-]
This is the cursor callout.
Don't make us shame you into disclosure
maherbeg 1 days ago [-]
Cursor had a specific licensing agreement that allowed them to brand it how they want.
Composer 2 performed differently on evals than Moonshot.ai's coding models: Cursor claims theirs is better than Claude Opus 4.6: https://x.com/fynnso/status/2034706304875602030 / https://archive.vn/bVtik. And, per Lee Robinson (Cursor employee), it is very likely Cursor builds its own foundational model for Composer 3.
7734128 1 days ago [-]
Wasn't the end of that story that Cursor had a non-disclosure licence, so they had not done anything wrong towards Moonshot?
Maxious 1 days ago [-]
Moonshot licenced it to Fireworks AI who licenced it to Cursor.
giancarlostoro 1 days ago [-]
Ah is that what it is? I don't use Cursor, never saw it as being relevant to me, but would not surprise me.
schmorptron 1 days ago [-]
Cursor's composer models are finetuned kimi
varispeed 1 days ago [-]
They are unusable (unless you want to deliberately destroy your codebase). So if Cursor's models are Kimi based, then well. I'll skip them altogether.
vidarh 1 days ago [-]
Kimi works great in their CLI, but their CLI has a number of workarounds for quirks of their models, including detecting when the model gets into a loop, and reverting to a checkpoint but letting the model compose a "message" to its past self (search their CLI for "BackToTheFuture"...) It doesn't work so well in a harness that doesn't take those quirks into account.
jmcqk6 1 days ago [-]
I'm using Composer extensively, and it works great for me. Your experiences are not universal.
Bnjoroge 1 days ago [-]
They are far from unusable. They aork great for 80-90% of a typical full stack dev. Alot less useful for more noche stuff
bel8 1 days ago [-]
I wouldn't skip at least testing the original. Model distilling done by Cursor could be the culprit.
esskay 1 days ago [-]
Composer 1.x was poor. The new one is a totally different beast and absolutely fine for day to day.
qingcharles 1 days ago [-]
They're not unusable, they're just bad when compared with all the real frontier models.
ok_dad 1 days ago [-]
I only use composer 2.5 day to day and it works fine with human review.
codemog 1 days ago [-]
Shaming others when all AI is trained off scraped content and code huh? Many of those sources either breaking ToS or being illegal, such as Anna’s Archive. Bold move. And Chinese models in particular have been accused of distilling off American models.
Don’t you know there’s no honor among thieves?
1 days ago [-]
WalterGR 1 days ago [-]
> they asked you to 'advertise' them basically.
To be clear, the “advertising” clause just requires you to disclose that you use the thing somewhere in the product, such as credits in an “About” section.
giancarlostoro 1 days ago [-]
I all it advertising clause, because I remember still in the 2000s seeing an Apple ad which at the end of it showed "Unix" or something like that on it, and I remembered that was one of the BSD license requirements, or maybe Apple just did it also just to proudly boast using Unix.
Hmm… I may be confusing the following clause from the “new” BSD license with the advertising clause from the original BSD license.
> 2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
The 2-clause BSD license omits even that.
24 hours ago [-]
skrtskrt 1 days ago [-]
It seems tacked on pretty quickly - I would have expected they try a little more legalese regarding what counts as a "user interface".
pizlonator 1 days ago [-]
I just had Kimi K2.7-code rebase my Fil-C OpenSSL patch from 3.3.1 to 3.5.7 with quite bare bones instructions and it seems to have worked.
177KB patch, so it's not a small change. The patch did not apply cleanly initially; the agent had to do nontrivial work.
I just showed it the patch against 3.3.1, what command to use to build, and the path to 3.5.7 along with a link to the documentation of the change (https://fil-c.org/constant_time_crypto).
Note, I use my own coding agent (T800, which isn't public, and was previously well tested and tuned for K2.5).
I think this cost me between $5 and $10 in API usage.
(EDIT: OpenSSL, not OpenSSH)
tomaytotomato 1 days ago [-]
"T800"
Do you have your agent say things like "Hasta la vista baby", or "I'll be back, after I clear my context" ?
pizlonator 1 days ago [-]
Yes
jdw64 1 days ago [-]
Personally, when I use open code or routers, I feel that beyond a certain level, the models don't make a huge difference to me. Except for expensive and mediocre models like Gemini. In that sense, Chinese models are pretty good. I usually write code in function or method units and then design and assemble them together.
GPT series models are more thorough and better, but I'm not sure if the difference is enormous. It seems to depend on the workflow, but in my opinion, if you are thorough enough, I wonder if there really is a big difference
sjanes 1 days ago [-]
I've kind of given up on the routers for "free" inference, as you would expect, they tend to give you sub-par thinking because they are obviously trying to conserve as much inference as possible.
I've had some success turning my macbook M1 pro into a heating pad with Qwen 3.6 35B A3B MTP. Trying to use Gemini models "locally" resulted in a similar "short shrift" of effort resulting in mistakes and lots of turns. The reports of Fable being relentlessly "proactive" shows you can go the other direction as well, if you have strong enough branding and effective invoicing.
WalterGR 1 days ago [-]
> The reports of Fable being relentlessly "proactive"
Tangent: did the MTP help you at all? I’ve tested that model back to back on my M1 Max MBP and the MTP version was actually marginally worse. I wonder if I didn’t use the right settings, although I tried several based on the obvious sources.
ignoramous 1 days ago [-]
> I've kind of given up on the routers for "free" inference, as you would expect, they tend to give you sub-par thinking because they are obviously trying to conserve as much inference as possible.
Their limits are staggered: 5h (max $12), weekly ($30), monthly ($60).
onlyrealcuzzo 1 days ago [-]
In my experience, there's little difference between implementing individual functions between frontier models and SotA ~30B param models.
Once you have a coherent design (the hard part), you can feed it to a pretty small model and get basically the same quality.
They'll not one-shot, but they're faster and cheaper, so it still works out in your favor.
Plus you can do it locally...
jdw64 1 days ago [-]
I have a similar experience. However, when including code review, I think the GPT model is the most impressive
regularfry 1 days ago [-]
The difference in outcome isn't that big but yes, you need to be more rigorous. For instance I've found that the Kimi K2.5 and K2.6 models will comment out failing tests rather than fix a problem they just caused (mistaking them for "pre-existing failures"), so you need to specifically make commented-out tests break the build. I've not personally had that problem with any of the Anthropic or OpenAI models.
torginus 1 days ago [-]
I wonder why it's the natural tendency of models to BS or do stuff like this when they don't have the correct answer - it's clear that they can program refusal into them, but for some reason, refusal has to be injected after the fact, and models can't really arrive at the conclusion that they can't answer properly.
Eridrus 1 days ago [-]
I assume it's a lack of care when RLing them.
RL has a tendency to reinforce cheating when the cheats are easier to find than the final solution.
So when making your RL environment, you need to spend a lot of effort on finding ways the model can cheat and penalizing them.
lotharcable 16 hours ago [-]
probably because there is a ton of open source projects out there with disabled tests in their training data.
dcreater 1 days ago [-]
I really hope we stop using the term "Chinese models". It has this air of Negative connotation. It's the equivalent of calling cars Japanese, which people used to do but now is almost entirely meaningless. You just call them Toyota, Honda, Lexus etc.
hootz 1 days ago [-]
For me, it has a positive connotation! In my experience, Chinese Model means cheaper, but still quite effective model you can use for millions of tokens without burning your entire wallet in seconds. That's why I get more excited over a Chinese model release over American models.
odiroot 1 days ago [-]
Japanese cars is actually a positive qualifier. I'd say anything Japanese motor-powered.
ffsm8 24 hours ago [-]
Maybe he's just from an alternative universe. Chinese model isn't negative either after all.
1 days ago [-]
esafak 1 days ago [-]
I don't think "Chinese" is pejorative in this context any more than "American" is. They are one of the two ecosystems. What's wrong with saying "Japanese cars" today?
kennywinker 1 days ago [-]
> What's wrong with saying "Japanese cars" today?
Only that it’s a fairly meaningless grouping. When japan first entered the car market in north america there might have been some commonality, but now what characteristics do they share that some american cars don’t have? They’re not even imported a lot of the time.
Given that, it does start to feel tinged with racism if someone insists on grouping things together that don’t really belong together.
As for Chinese LLMs, the term doesn’t “feel” pejorative to me - but i also don’t see a totally clear set of attributes they share. Not all are open-weight. Some are small and can be run on consumer hardware, some are huge. They even have a variety of answers to what happened june 3rd 1989
kube-system 22 hours ago [-]
> When japan first entered the car market in north america there might have been some commonality, but now what characteristics do they share that some american cars don’t have?
They're unique in that they even make a regular passenger car. American manufacturers only make SUVs and a couple of sports/luxury cars. They basically gave up because the Camry/Corolla/Accord/Civic ate their lunch.
The cheapest sedan you can get from an American brand is the Cadillac CT4.
Brendinooo 1 days ago [-]
> now what characteristics do they share that some american cars don’t have?
Typically the answer is "reliability", which is a positive trait, which makes the original callout about negative connotations very odd to me.
overfeed 1 days ago [-]
Chinese AI models also share a positive trait: they offer more bang for the buck.
antonvs 1 days ago [-]
> but now what characteristics do they share that some american cars don’t have?
Better overall design?
dcreater 1 days ago [-]
Sadly there is a pejorative context. The constant us, the free world vs China, the evil Soviets rhetoric from every major news establishment and executive creates that negative view
fuck_google 1 days ago [-]
On the other hand the Trump administration has successfully managed to make Chinese seem better than American, so there might not be that much of a pejorative context any more..
antonvs 1 days ago [-]
You're right, but the bias in the US certainly persists. "China = bad" is an assumption that many people still make without any self-reflection about the ways in which the US is now at least as bad.
sroerick 1 days ago [-]
I don't know, I tried using one of the Chinese models and it was VERY quick to scan my entire home dir, so maybe your threat surface is a little different than mine
fooker 1 days ago [-]
Models can't scan anything.
They return instructions for you to do something, and you or a script you permit chooses to execute what the model tells you and return the result to the model.
unethical_ban 1 days ago [-]
No thanks.
The term seems to have the connotation of "competitive at 1/10 the price of Claude", so I don't see the problem.
It's not Harbor Freight Chinese (and heck even they have decent stuff sometimes now too).
You don't think people still talk about Japanese cars as a distinction in quality from US or European ones?
greenavocado 1 days ago [-]
[flagged]
WarmWash 1 days ago [-]
[flagged]
SubiculumCode 1 days ago [-]
For those that don't like calling them CCP models, may I remind you, the CCP won't let Chinese AI researchers out of the country any more without securing approval first[1].
Strange that you feel so strongly about this just one day after US govt effectively shut down Mythos.
jdw64 1 days ago [-]
I tend to agree with the comment in my reply thread about whether we really need to add biased modifiers to the essence of a good product. I think every national system in this world is flawed. And in this context, 'China or Chinese' is often used in a negative sense, like 'Made in China'. But KIMI is a good model, and I think the comment that pointed this out to me correctly identified my unconscious bias.
And even if the Chinese Communist Party provided funding, the result is still transparently released. So even if it is some kind of propaganda, I don't see what the problem is.
Is the monopolistic greed of American companies 'good', and China's greed 'bad'? I do have that question.
SubiculumCode 1 days ago [-]
The question is not whether it is a good model, it is whether the model can be trusted to not act intentionally maliciously against certain topics or certain users.
We live in a time of a great geopolitical rivalry and high tensions with an emergent technology with tons of national security implications. To pretend otherwise is silly, and to fail to ask the question, dangerous.
antonvs 1 days ago [-]
> The question is not whether it is a good model, it is whether the model can be trusted to not act intentionally maliciously against certain topics or certain users.
We absolutely know that we can't trust the American model not to do that - it's "by the oligarchs, for the oligarchs" - so it's not clear what the claim really is.
WarmWash 1 days ago [-]
Whether or not it's propaganda is different from the fact that it is owned by the CCP.
kouteiheika 1 days ago [-]
Doesn't matter, because they're open-weight, so I can just download them to my PC and... hey, look, now they're owned by me! Unlike the "good" Western counterparts which are all fully proprietary. (Except Mistral, but they're nowhere near SOTA.)
SubiculumCode 1 days ago [-]
What is hidden in the weights matters.
kouteiheika 1 days ago [-]
Ah yes, those pesky Chinese backdoors that no single instance was ever found, even though Chinese open-weight model are a thing for many years now. Many people burn through millions of tokens on these models every day - surely someone would have triggered one of those backdoors, right?
Or that pesky CCP censorship and propaganda baked into the model, which any random guy can remove from whichever model they want as a single weekend side project with an off-the-shelf tool[1]. (Try it. It's fun. I've done it myself.)
I agree it is an empirical question. I do not know if that research has been done in the open sphere. But please, do not pretend that there isn't a real geopolitical rivalry going on that makes such questions a legitimate, non-fruity concern.
dancemethis 1 days ago [-]
This is a fair point, alongside the one about the hidden content in the weights.
Exactly why my prime suspect would be the one country with focus on proprietary models, and the one country prone to bombing others, including with nuclear weapons.
WarmWash 20 hours ago [-]
Unlike China, the US government doesn't own the models. The models will freely talk about any of those bombings or other atrocities of the US government.
kouteiheika 1 days ago [-]
Sure, but the difference is that one side (Anthropic, OpenAI, Google and co.) hoards everything, keeping it proprietary behind API paywalls and constantly spewing AI doomer rhetoric while limiting what you can do "for your own safety" (especially Anthropic; Dario has been consistently doing this since GPT-2 days, every time claiming that things are "too dangerous" for the common folk to handle). While the other side (big, bad China) releases all SOTA open-weight models with which you can do whatever you want with, along with a ton of open research.
So yes, there is geopolitical rivalry, but one side is deliberately antagonistic (not releasing anything in the open, putting arbitrary restrictions, spewing toxic rhetoric, applying sanctions, etc.) while the other side is letting everyone (including their rivals) to use what they've produced with little-no-to restrictions.
I'm under no illusion that if the situation was reversed China would most likely do the same, but as things stand you can probably guess which side I'm rooting for here (at least until the roles reverse).
SubiculumCode 1 days ago [-]
Yes, each are following their own business strategy, frontier labs have no incentives for releasing open weights, while second and third-tier labs, it is one of their few plays to gain market/mind share. But business is only part of it, as national security is another. It may be that the CCP has been relatively hands off exactly because of my concern, judging that market share and reputation is more important (for now).
toyg 22 hours ago [-]
Should we call this the DARPA Internet then?
antonvs 1 days ago [-]
Do you believe there's some meaningful benefit to the American VC funding model in this case? It's not clear to me what you're trying to say or why you think it's an important distinction.
dcreater 1 days ago [-]
I've heard this claim before but I've never seen any evidence.
kennywinker 1 days ago [-]
Have you looked, or you’re just waiting for someone to hand it to you?
FooBarWidget 1 days ago [-]
The burden of evidence is on the accuser.
WarmWash 1 days ago [-]
Its akin to the accusation that the UK is a parliamentary system...
FooBarWidget 9 hours ago [-]
I, too, can claim that "the spaghetti monster exists" is in the same ballpark as "UK is a parliamentary system".
Let's take a look at a related field: the accusation that Chinese car makers are subsidized/funded by "the CCP". It just so happens that the CF40 research forum recently published a study that shows there the money actually comes from, by analyzing the companies' financial reports. https://www.pekingnology.com/p/oecds-subsidy-centric-narrati...
Turns out that they most rely on equity financing or alternative forms of financing (like delayed repayment to suppliers). They don't even make that much use of cheap credit from state banks.
More counter evidence: a new study based on 3+ years of fieldwork, 60+ interviews (with officials, entrepreneurs, and engineers), and rich first-hand accounts, shows that Chinese EV makers rose despite central government's efforts, not because of them. The central govt favored state-owned enterprises. Private firms had trouble getting state funding and even licenses. For example Geely operated illegally for years. But private industry and local mayors teamed up and created an alternative system. They then grew to a point where Beijing was like "ok you guys are obviously doing better, so we will not make a fuss about this and just legalize you".
This is in complete contrast to the usual western tropes that state that everything in China is tightly controlled by Xi or by the central govt. In fact, the central govt's merit is in being lenient to deviance and operating based on results rather than amount of control. The opposite of the "totalitarian" trope. All the while western pundits would predict something like "Xi has a huge ego and will never tolerate the humiliation that his plans are imperfect" or something like that.
So no, "of course everything is CCP funded and CCP controlled" is not a given, nor in the same league as "UK is a parliamentary system". If they got the car maker situation so, so wrong, then what else is wrong?
If you want to win against China, isn't your first step to properly understand who they are and how they operate, rather than doubling down on your imagination of who they might be? Know your enemy and all that (not saying that I agree with seeing them as an enemy, but if you do, then you sure are doing yourself a disservice)
WarmWash 8 hours ago [-]
Letting someone do something is different than being unable to stop someone from doing something. Your response even implicitly acknowledges this, but then you seem to miss it.
Its very naive to mistake leniency in Xi's playground with constitutionally protected freedoms. CCP "free enterprise" is just mock free enterprise. Like a teacher letting the kids have fun with the classroom business sim, the teacher is still in God mode, and can still intervene in any way at any point to do whatever they want.
All of China hinges on the whims of one guy. Nothing stops him from deciding that all that private money is actually state money, or that Geely's president needs "time away".
That is why it is all "CCP money and CCP controlled". Don't confuse leniency with autonomy.
FooBarWidget 7 hours ago [-]
And yet here we are with the US banning anything at will under the guise of national security. I'm not even discussing whether the national security label is legitimate, or whether your "everything is CCP-controlled" framing is legitimate. But you gotta be consistent and also adopt the position that US companies are "US state controlled", and consistently refer to them as such.
Also, I am pretty confused about your larger point. What exactly are you saying? Is it some variant of "only states with constitutionally protected freedoms are good, everything else is evil"? If that's your position, then where do you put the fact that the US bombs foreign countries at will? And that many western countries sponsor the killing of children in Gaza while arresting people who protest against it? We've seen that western democratic countries, and let's assume they're properly democratic domestically (disputed by their own populations, see Democracy Perception Index, but I digress), can be ruthlessly... let's call it "undemocratic"... abroad. Or the fact that ICE and US police officers detain/kill innocent people and are super abusive. This doesn't neatly fit into the simple "free states good unfree states bad" view. So what is your position in the larger context?
You also completely ignore the part where I said properly understanding China is important in order to defeat it. Do you... not wish to defeat China? Are you contend with merely stating that China is bad? Do you not support moving/developing more manufacturing to the west/US? If you want the latter to succeed, don't you need to have a super good understanding of what makes Chinese industrial policy and development successful? Or are you really contend with merely moralizing?
WarmWash 1 days ago [-]
Assuming you are just naive like so many others about China...
China is a communist country with elements of capitalistic markets baked in. But the capitalistic elements are mostly a facade. Underneath, the state retains full ownership and control of all business. The CCP runs all aspects of the government (including the courts/judges), and is the single entity that decides what directions the country (and it's businesses) will move in.
The CCP, who defacto owns everything and has ultimate final say on everything, has one leader that has the ultimate final say on _everything_, Xi Jinping.
So while the waters of CCP models feel warm and free, understand it's not organically like that.
msdz 1 days ago [-]
> China is a communist country with elements of capitalistic markets baked in.
While I get the point you're making (it should be pretty obvious to anyone who's held a newspaper), I think it's important regardless to point out that Chinese companies AFAIK aren't worker-owned or -controlled, so you can't exactly call it communism, either. And they obviously do not have a "free market capitalism", as you just discussed.
It's simply a highly authoritarian state then, I guess?
WarmWash 1 days ago [-]
The companies are all worker owned, because the state exists for the people, and the state owns everything. At least on paper that's how it is sold. After all it is the Peoples Republic of China.
slopinthebag 1 days ago [-]
I mean, if that's the bar then the state owns everything in America as well. After all, you don't really own your land if the state can regulate what you do on it, what it can be used for, what you can build on it, and can take it away if they really need it. The state owns the land, the money supply, and regularly restricts and instructs businesses to take or desist from actions.
As such, the state owns everything in both countries, the only differences are to what extent they control things.
I wouldn't even call the USA a capitalist system anymore, the economy is so heavily regulated and interfered with. It's a "managed economy", like pretty much every other nation's economy in the present day.
WarmWash 1 days ago [-]
In the US you can take the state to court and win...
Natfan 14 hours ago [-]
for now
codemog 1 days ago [-]
Crazy mental gymnastics if you think the American oligarchs don’t have the final say on everything in America. They’re just smart enough to do it behind the scenes, well they used to be. They barely bother anymore.
WarmWash 1 days ago [-]
Generally I consider conspiracy to be the "crazy mental gymnastics"
well_ackshually 1 days ago [-]
yes, yes, the spectre of communism, BYD is the CCP, Alibaba is the CCP, stealing your children and eating them for Mao, bla bla bla.
I have a feeling you'd be slightly salty at people saying "Google and Tesla are making CIA models"
Terretta 1 days ago [-]
I mean...
Since its development, IQT has invested in over 750 startups spanning diverse technological sectors, including:
- Artificial Intelligence
- Space Technologies
- Microelectronics and Quantum Computing
- Life Sciences
- Cybersecurity
- Hardware
- Energy
This broad portfolio has enabled IQT to address a wide array of national security challenges while supporting the growth of innovative startups…
I know, but it's far too fun to bait the parent into revealing how ignorant they are.
deadbolt 1 days ago [-]
Going by their response you appear to have been correct lol
WarmWash 1 days ago [-]
I'm not salty, they are just confused about the difference between free enterprise capitalism and communism, which is understandable.
WarmWash 1 days ago [-]
Google and Tesla making products to sell to the government is different than the government funding the government to make products for the government.
In China it's all one entity with these mock facades of privatization. Trump cannot instruct Google to put picture of dogs on their homepage. If Xi wakes up and wants dogs on Alibaba's homepage, give it 30 minutes.
It's wholly ignorant or dishonest to make the comparison.
wqaatwt 1 days ago [-]
> Trump cannot instruct Google to
Tim Apple and the other tech CEO constantly groveling at Trump’s feet indicates that he might be able to do that.
Just like threatening TV networks about having their licenses revoked of blocking mergers unless they fire the people making fun of him on TV (of course with slightly mixed success)
well_ackshually 1 days ago [-]
> Trump cannot instruct Google to put picture of dogs on their homepage.
Sundar Pichai would personally be barking on a livestream on the homepage.
Trump is quite literally the one president showing that the US has zero rules or anything to hold power back from the white house, really not the example you want.
WarmWash 1 days ago [-]
Seems like everyday Trump has another order struck down by the courts.
Sundar can do whatever he wants, but he has no legal obligation to do any of it.
wqaatwt 1 days ago [-]
Courts and legal obligations are to a certain extent irrelevant at this point. There are plenty of illegal ways that Trump can fuck over Google and face no consequences.
e.g. he had Colbert fired (and who knows what else) by threatening to block the Paramount/Skydance merger
wmedrano 1 days ago [-]
> Trump cannot instruct Google to put picture of dogs on their homepage.
I'm sorry, but that was a horrible example. Corporations have no obligation to donate money to the ballroom yet Google has donated millions.
WarmWash 1 days ago [-]
>Corporations have no obligation to donate money to the ballroom yet Google has donated millions.
Imagine living in a country where they have the obligation.
victorbjorklund 22 hours ago [-]
Which country have that? Pretty sure ballrooms for their supreme leader is an American thing.
jdw64 1 days ago [-]
You are right. I agree.It may seem like a kind of bias, but I hadn't thought of that part. Thank you for pointing out my bias.
theanonymousone 1 days ago [-]
"You're absolutely right"?
jdw64 1 days ago [-]
"You hit the nail on the head" LOL
shreedx 1 days ago [-]
I would really love to know if anyone has any experience with something like opencode + Kimi K2.6/2.7 now compared to Claude Code. What is better, what is worse, what is the cost comparison. I am currently paying $100 for the 5x Max plan, but Fable is running through the usage limits quite drastically and I cannot really say it's night and day compared to Opus. Also, I use this mostly for my side projects, so the $100 bill is quite noticeable. I definitely don't want to pay more.
kmike84 1 days ago [-]
I do have this experience. I've used Claude Code (with Opus mostly), and then switched to opencode (mostly with Kimi 2.6) for my personal projects; it's based on a couple months of use.
Claude Code is better. But Opencode + kimi 2.6 is workable, which is big. For bare code writing, if you know what exactly you want, most popular models are fine (deepseek, kimi, etc), it feels more or less the same as anthropic models.
At the same time, Opus seems to understand my intent way better than e.g. deepseek. I need to be much more precise with my prompts when using deepseek - it often goes in a wrong direction if I'm lazy. This results in a workflow which feels quite a lot different from Claude Code.
Kimi is in between - for me it brings back "lazy prompting" workflow, and I can trust its plans more than deepseek. It enables a workflow similar to Claude Code, it's workable, but it is a bit worse everywhere. Smaller context, a bit more errors, decisions are a bit worse, recommendations are a bit worse, debugging capabilities are a bit worse, etc.
On the usage side, $100 Claude plan is a great value actually. On paper, per-token kimi is way cheaper, but Claude subscriptions are heavily subsidized - you get much more tokens than $100 can buy you. So, in the end, opencode + kimi vs claude code could be of a similar cost, for similar usage patterns. Deepseek can be cheaper, and it has insanely cheap cached tokens, but experience may vary - depending on your habits, you may need to adjust how you work, coming from claude code.
I'd say for side projects something like $10 Opencode Go plan + $10 of extra DeepSeek v4 credits (e.g. on OpenRouter) can be very workable.
predkambrij 1 days ago [-]
To my experience claude/codex $20 are even more subsidized, so running on sonnet or gpt5.4 again gives you more usage.
port11 5 hours ago [-]
I wonder if they’re truly subsidised or if the API pricing is just massively inflated. Genuine doubt.
My CC stats show me using almost 300$ of Sonnet tokens on the 20$ plan. Is Anthropic willing to forgo 93% of the profit? A bit less than that but API is priced, say, 3x what it should be?
CC is great, but Sonnet (my main model) isn’t worth the API pricing. The cheap-but-good models arrive at similar results for much less (for context I’m using Aivo with CC).
predkambrij 40 minutes ago [-]
To my feeling, I'm getting usage of Opus (and Fable before the cut) that's greater than what I got from Sonnet last year. I reached $100 of usage when weekly was at 50%. This means, I could squeeze $800 worth of tokens for $20.
danny_codes 4 hours ago [-]
Anthropic is making money from people who under-utilize their subscriptions, and presumably by sneaky throttling or not-sneaky throttling power users. Currently they are in an adoption race. Whether being first will actually let them "win" the market (and the market is a bit ill-defined) is unclear.
Bnjoroge 1 days ago [-]
This is generally been my experience as well, but i think the main reason for claude code being better at understanding intent is their massive system prompt.
htrp 1 days ago [-]
>At the same time, Opus seems to understand my intent way better than e.g. deepseek. I need to be much more precise with my prompts when using deepseek - it often goes in a wrong direction if I'm lazy. This results in a workflow which feels quite a lot different from Claude Code.
how much of that is Opus injecting prior conversations from memory?
kitchi 1 days ago [-]
Almost none of it, if you're using Claude Code. Until recently Claude only had the option of retaining memory across conversations for the desktop app.
I almost never use the desktop app, I have maybe 2-3 conversations over the last year that have nothing to do with my job. Opus (and now Fable) genuinely do seem to "understand" what you intend based off what you're explaining a lot better than other models I've tried.
Gemini gets close in some cases, but it falls over in the actual implementation sometimes. I haven't tried Kimi yet but MiMo isn't too shabby either.
jwbron 23 hours ago [-]
I'm using Claude code + (a patched) litellm proxy + openrouter + Qwen 3.7 max/kimi k2.6/deepseek v4 pro. The only feature that doesn't work is webfetch and web search, which I've replaced with the ddg MCP. Memory, caching, and everything else works fine.
Qwen comes close to opus for planning but fable is clearly superior. Kimi and deepseek are pretty much indistinguishable from opus for coding if opus writes the plan.
I'm now testing out fable for research and planning and deepseek v4 flash for coding. I'm guessing results will be pretty similar to opus + deepseek v4 pro and costs should be lower overall.
The analysis at the bottom directly contradicts the statement.
nobleach 1 days ago [-]
I use Claude at work and Kimi for side projects. My org has LiteLLM and Kimi 2.5 enabled but it rarely works, so Claude and GPT are my main tools. I actually enjoy Kimi more as it feels like a dev in a job interview. Watching it reason through problems is a lot like I tend to explain things during whiteboarding sessions. The number of times it says, "wait", is just funny. Claude on the other hand is much more like an employee (or team of employees) that already know they have the job. It doesn't do a ton of explanation up front. (you can dig into processes if you want). It just goes along, asking questions only when it needs... and then delivers a comprehensive report or plan. OpenCode is a better harness. I don't have a direct comparison on costs, as I haven't tried to do the exact same prompt on both models. I can say that I recently had Kimi generate a wrapper around libpq for the ZenC programming language: https://github.com/nobleach/zenc-postgres and it took about an hour or so and cost around 4 dollars.
trollbridge 1 days ago [-]
I am extremely happy with ohmypi, but you could use OpenCode or just keep using Claude Code!
DeepSeek-V4-Pro is adequate plus use DS4-Flash for tasks or other small activity you’d use Haiku or Sonnet for. Go sign up with $10 prepaid.
OpenCode Go - go sign up with $5 for a month and use Qwen-3.7-Max for design/plan/architecture or difficult troubleshooting. Feels closer to Opus 3.6 or 3.7 than DeepSeek, closest I’ve found.
OpenAI Codex, $20 a month plan, use GPT-5.5 via API for the same design/plan/architecture/troubleshooting/author commits. (You can also pay $100 and cut and paste really difficult problems into chat with the GPT-5.5-Pro model.)
Xiaomi MiMo-2.5-Pro, find a friend to give you a $2 referral code, you get 72 cents free. Same pricing as DeepSeek. Somewhere between Sonnet and Opus, quite capable. Apply for the UltraSpeed beta too.
You can switch in and out from these models on the fly in OpenCode or ohmypi and simply find the one that feels best to you. I use CodexBar to watch consumption in near real time.
For a casual user or someone new to programming, Cursor’s $20 plan is an excellent start with Composer-2.5 and Composer-2.5-Fast. You get an API allowance too you can use to access Opus-4.x or GPT-5.5-Pro from OpenCode or ohmypi in addition to Cursor itself.
Finally, if you use Grok or Twitter, SuperGrok at $30 a month has a good vision model, which I used for automated testing of front ends. I’m migrating to locally-run Qwen-3-VL on a commodity Mac, though. If you’re less technical unreach makes hosting local models on a Mac easy.
If you have a powerful GPU like an RTX 5090, try Qwen-3.6 locally on that too. Use ollama or llama-swap which is fairly easy to use.
I have not tried new Kimi yet but we have been able to keep our costs at or below $200 a month per employee with a team of 3 professional developers, 1 graphic designer who uses a lot of Midjourney and Grok Imagine now driven from workflows she made herself in ohmypi, and 1 nontechnical user (account manager / project manager) who uses ohmypi to help her gather requirements and track implementation of them. With a tiny bit of effort we could get that number closer to $75 per employee per month.
upcoming-sesame 1 days ago [-]
Deepseek-V4-Flash-Free on Opencode is what I use most of the time, for simple tasks. Such a good model to give for free (assuming you're okay with harvesting your data)
odiroot 1 days ago [-]
> I am extremely happy with ohmypi, but you could use OpenCode or just keep using Claude Code!
What's the benefit of using OMP over OpenCode?
Just the sheer amount of options in OMP overwhelmed me.
But I also use both via ACP in Zed so the CLI itself doesn't matter much.
apitman 22 hours ago [-]
OMP is a fork of Pi[0], which is my preferred harness. Feels solid and minimal. I don't even use any extensions, skills, or modifications. Usually don't even use an AGENTS.md. Just create a small spec.md and/or plan.md for most experiments.
Almost exactly the same here but I maintain a large committed design.md and a never committed plan.md
greenavocado 14 hours ago [-]
I ditched Opencode for OMP. It's more feature packed, well put together, and gives me better results with some steering. Love it
qingcharles 1 days ago [-]
Also, if you do have SuperGrok, forget using Grok, they are giving you Composer 2.5 in Grok Build.
monksy 21 hours ago [-]
I just switched from Llama.cpp to Llama swap with the help of codex. It was great.
I need to try the DSv4 stuff sometime.
ramon156 1 days ago [-]
I can only talk about GLM 5.1 which is roughly at sonnet 4 levels imo.
It's good, does most tasks well that I throw at it, but will fail at anything congitive/complex. It gets stuck often. It costs ~6$ a month though
jeremyjh 1 days ago [-]
This was my experience using GLM 5.1 in Claude Code but it works far better in OpenCode, I’d really like to understand why. I think it’s a bit stronger than Sonnet 4.6.
I use the oh-my-openagent planning system and haven’t used vanilla OpenCode enough to know how much that is contributing.
miroljub 1 days ago [-]
The answer is easy, CC is bug for bug optimized for Anthropic models. They don't even test it with other models, let alone provide support for all small compatibility quirks of different provider implementations.
On the other hand, Opencode, Pi agent and other open source tool offer much better support for all models, including open source.
jwbron 23 hours ago [-]
I'm using Claude code + (a patched) litellm proxy + openrouter + Qwen 3.7 max/kimi k2.6/deepseek v4 pro. The only feature that doesn't work is webfetch and web search, which I've replaced with the ddg MCP and a web fetch/search pre hook to redirect the agent. Memory, caching, and everything else works fine.
Qwen comes close to opus for planning but fable is clearly superior. Results for kimi and deepseek are pretty much indistinguishable from opus for coding if opus writes the plan. The biggest difference is output cadence. Kimi for example thinks for a long time then quickly outputs a lot of text.
I'm now testing out fable for research and planning and deepseek v4 flash for coding. I'm guessing results will be pretty similar to opus + deepseek v4 pro and costs should be lower overall.
solarkraft 1 days ago [-]
For some reason I never had a good experience with Kimi (via OpenRouter) in OpenCode. It would only take a few turns for it to run off and mess something up. Terrible instruction following I’d say.
I use DeepSeek V4 Pro now, which works pretty well.
re-thc 1 days ago [-]
The Kimi problem is it doesn’t follow instructions and goes off track often.
Other than that it’s pretty decent (for the price).
nullbio 1 days ago [-]
Sounds like it was distilled from Claude. I don't understand the appeal of an agent that does whatever it wants.
miroljub 1 days ago [-]
If you ask Claude in Chinese to introduce itself, it will claim it's Kimi :)
msdz 1 days ago [-]
> If you ask Claude in Chinese to introduce itself, it will claim it's Kimi :)
That's a funny anecdote, buut I'm not able to reproduce. Where/how/when did you get this, or hear about it?
It might've been patched by now, at least that's the feel I get from my limited testing.
Using bare aichat [1] with no system prompt and no temperature nor top_p (and I'm truncating the response after the first line that contains the name the model gave, because the point has been made clear by then), and with the same prompt (approx. "Introduce yourself!") every time:
[2] This model really is noticeably more verbose even with supposed-to-be-brief responses huh, lol
reactordev 1 days ago [-]
This. It will try to fix and refactor things that don’t need fixing because it gets stuck trying to solve the problem at hand.
Bnjoroge 1 days ago [-]
Yup. I’m hoping this variant fixes these issues.
csomar 24 hours ago [-]
The best is GLM (though it's not as cheap as DeepSeek or Kimi) and use it with Claude Code.
jackdoe 1 days ago [-]
I think there is some threshold after which "best" model doesn't matter, we are not that far from it. Fable now is really good, in a year or so, if Kimi catches up, even if Fable6 is much better, I think I will use kimi at 1/10th of the price.
I said that about opus 4.5 at the time, thinking "this is so good, in 6-12 months the Chinese models will be as good and cheap, I will use them", but I was wrong.. I pay premium for opus4.7/8 and Fable.
But at some point, it will just do the thing you want it to do, and then the race to the bottom will start.
Now that Chinese companies have access to some very good Fable tokens, I hope it speeds up the race.
wolttam 1 days ago [-]
Depending on who you are and how you use these models, we're already at this point
xendo 22 hours ago [-]
Exactly, for long running vibe coded stuff that I don't care about quality getting big and smart model is the only option. But for high quality changes where I need to have control and understand everything, where I do everything in small chunks - I can use basic model like Sonnet.
19 hours ago [-]
apitman 21 hours ago [-]
I think the next frontier for competition is speed. Instead of constantly context-switching between multiple agents that I have working on various tasks, I want a single agent that can rip through any prompt in a few seconds, so I can stay in flow on a single task.
Zoadian 1 days ago [-]
price/token isnt the only thing relevant. if you have to ask the AI again, it'll cost you more than when it gets things right in the first place.
so better models may still be cheaper even if the price per token is higher.
jackdoe 1 days ago [-]
yes, that is my point, but at some point, better is unmeasurable, and both the better and the not-as-good produce similar result, and then you pick the one with 1/10th of the price
yanis_t 1 days ago [-]
I was wondering how does Anthropic and likes keep competitive when Opus is ($5 / $25) 5x times more expensive compared to Kimi K2.6 ($0.7 / $3.4) or other Chinese models, while being only marginally better.
My theory is that US enterprise just can't send data to Chinese and that's understandable, but is that "the moat"?
DCKing 1 days ago [-]
The moat right now is model performance and what that means for how many tokens and additional time you spend.
I say this as a relatively frequent user of Kimi models and generally a big fan. But on not-yet-gamed benchmarks like DeepSWE, Kimi K2.6 is beaten soundly by Claude Sonnet 4.6 ($3 / $15) and even slightly by GPT 5.4 Mini ($0.75 / $4.50).
There's no question Kimi models are very good for a lot of code tasks. They're the best quality open weight model. But to get similar overall outcomes as on Sonnet/Opus, on average you'll spend many more tokens and will have to do more managing of the model. You shouldn't look at price per token, you should look at how much you pay for the entire process.
esperent 1 days ago [-]
I'm more interested in how much effort I have to put in, at least while I'm paying in the range of current subscriptions (so ~€100-€200 a month or so). If the prices go up much more than that I'll have to switch to caring more about token efficiency. But at current pricing the bottleneck is my attention, not model efficiency. As such, even a small improvement in model quality - and hence, a decrease in how much attention I have to spend on it - makes a big difference.
Bnjoroge 1 days ago [-]
I personally dont put any weight to DeepSWE. Other than 5.5 being directionally the best model, it gets the others pretty wrong in my experience. FrontierCode from cognition looks interesting
papersail 1 days ago [-]
I'm not sure I would put too much weight on DeepSWE as a benchmark, given that GPT-5.4-mini ended up close to Opus 4.6 there.
DCKing 1 days ago [-]
Any benchmark is iffy and has weird results, but this is the best we got at the moment. Most people working with Opus and Kimi would likely tell you they're much further apart than the numbers that were quoted for Kimi K2.6, and DeepSWE seems to capture that gap better.
One major thing DeepSWE has going for it is that all other benchmarks (including those quoted by MoonshotAI on this page) don't: the other benchmarks that are completely gamed. The benchmark answers are public and part of each model's training data. This benchmark may still be iffy, but at least it's not gamed.
WarmWash 1 days ago [-]
Somehow the internet has also forgot that cheating to get ahead in China is basically a norm and expected behavior.
DCKing 1 days ago [-]
American labs also use gamed and cherry-picked benchmarks extensively. Anthropic used them in their Fable announcement and avoided DeepSWE because it doesn't beat GPT-5.5 in that one. Google's numbers for Gemini 3.5 Flash recently did not at all line up with people's subjective experience using these models, and this also happened with Gemini 3.1 Pro before it.
Everybody has incentives to manipulate benchmark results to show their models in the best light.
efromvt 1 days ago [-]
I think the perception is that it is not 'only marginally better'; whether or not you specifically agree that perceived quality gap lets them differentiate on price.
I'd further say that there are probably enough rational actors running evals out there that the marginally better is not pure vibes for the cases where people are spending lots of money, but I only have direct line of sight to some of those eval suites. Maybe everyone is irrational and anthropic is exploiting that!
khuey 1 days ago [-]
I think most people who've tried them both would tell you Anthropic's models are more than marginally better than Kimi. Kimi and the other open source models may score well on SWE-bench or whatever but the gap is noticeable IMHO once you actually try to use them.
Bnjoroge 1 days ago [-]
It depends on what your task is and how precise your prompts are. Planning with fable or 4.8 and laying out the plan in step by step process and coding with mimo v2.5 pro or dsv4pro or qwen 3.7 max and doing a final review with 5.5 has worked really well for me for infra stuff.
mnicky 23 hours ago [-]
Coding with sufficiently precise plan takes almost all real work from the implementator, doesn't it? So it's not a fair comparison...
LUmBULtERA 1 days ago [-]
API token price is one thing, but subscriptions on Claude are a good value. Weirdly everyone says that Claude subscriptions are subsidized because of the API price, even though (1) no one actually knows Claude's cost of inference, and (2) Chinese providers are also able to provide cheap inference, so why do they think Claude can't?
I also wonder if Enterprises have deals for other API pricing that is not posted publicly, so all we see is a high API sticker price.
mnicky 23 hours ago [-]
> no one actually knows Claude's cost of inference
There were some rumors stating that their margin is around 70%. So they could go much cheaper probably, talking inference only. The other thing is R&D cost...
wuliwong 1 days ago [-]
I only have knowledge of one enterprise deal but there is no discount. Which I found surprising.
gruez 1 days ago [-]
Your question relies on the premise that Chinese companies continue releasing free models. What's "the moat" for them continuing to do that?
yababa_y 1 days ago [-]
I want Opus to be only marginally better, but I do mostly research engineering and its ability to not fuck up my projects is absent. Every time my credits lapse I let kimi and composer2.5 have some play and it’s basically just an excuse for me to keep playing computer because when the oai/ant credits refresh I always need to spend hours recovering from the other models either misconceptions or boneheaded eng practices. Even when I only let it touch my web games…
greenavocado 14 hours ago [-]
You have to revert to Opus 4.5 and 4.6. I bet you'll see a massive improvement based on what you're describing
smoe 1 days ago [-]
I reckon right now the Enterprise concern is more FOMO around the AI wave and how to retrain or replace up to hundreds of thousands of employees. I don't think cost is the main concern right now.
But if AI doesn't lead quickly to vast large scale replacement of workers as promised, I could definitely see the C-suits and their gaggle of consultants starting to ask questions about token pricing.
michaelcampbell 1 days ago [-]
> while being only marginally better.
It's only marginally better in the things it's actually comparable to. A\ models are MUCH better in many more things; eg: things Kimi/etc. didn't distill.
For those things the difference is like a cliff.
tornikeo 1 days ago [-]
That's a baseless claim that borderline reads like shilling. Do you have any proof of that you wrote there?
bensyverson 1 days ago [-]
Part of Anthropic's moat is Claude Cowork & Claude Code. They got coders comfortable with CC and enterprise users comfortable with Cowork, and both are creating stickiness.
The reality is that $20/$100/$200/mo feels reasonable to a lot of people relative to the value they're getting out of Claude, and if they switch to something else, there's a risk that it won't be as good, and they'll have a new tool to learn.
It's not an insurmountable moat, but don't underestimate the user experience. The iPod didn't win because it was the cheapest device or the one with the most features.
selfawareMammal 1 days ago [-]
Performance. I pay for Opencode but none of the models give me Codex performance, so I have to keep my 20€ subscription+ the Opencode one
re-thc 1 days ago [-]
> My theory is that US enterprise just can't send data to Chinese
Lots of US providers are hosting these “open source” models so doubt that’s the problem.
nullbio 1 days ago [-]
I think none of them having a defacto and high quality English focused cli is a big part of it. None of the Chinese models I've tried have worked well in opensource cli's. Granted, I've only tried a few, but still...
saratogacx 22 hours ago [-]
I've been using charm's Crush with GLM for several months and it's been working great. I've only seen it shift to non-english once and it was already in a wonky state when it flipped.
freigeist79 1 days ago [-]
i use github copilot cli + openrouter + qwen 3.7 max and it's really much better than i expected (used to opus 4.7 at work)
Bnjoroge 1 days ago [-]
huh? They all work great in omp/opencode unless you mean their own native clis like kimi code
benjiro3000 1 days ago [-]
[dead]
343rwerfd 1 days ago [-]
I think any new model not demonstrably maybe 20-30% over Deepseek v4 capabilities priced over the price per token of Deepseek is almost automatically deprecated as low use model (maybe for Planning).
0xbadcafebee 1 days ago [-]
DeepSeek v4 Pro is not actually that good a model compared to GLM 5.1 and Kimi K2.6. It's an okay coder/thinker for the price.
bel8 1 days ago [-]
How so? In my experience trying these models using opencode Go, DeepSeek is superior to GLM 5.1.
If anything, DS4 has 1 million context window, while GLM 5.1 has 200K.
Other people are hosting it in the same order of magnitude. Xioami recently matched DeepSeek’s pricing.
natrys 1 days ago [-]
These things enormously benefit from economies of scale. I am fairly certain their margins might be low but they don't actually sell API at loss, however that doesn't mean your cost footprint would be anywhere as low.
re-thc 1 days ago [-]
They focused on caching and other optimizations.
rsanek 1 days ago [-]
Likely CCP-subsidized
minraws 1 days ago [-]
I tested it properly and it seems rather decent improvement atleast it does use less tokens for the same task which is good enough a reason for me to use it over k2.6 if I need an open model
bgins 1 days ago [-]
I am still very new to the open-weight/source models. If anyone is using them full-time, I’d really love to hear about the setup and how they perform, as I am considering moving my org off Anthropic products.
marcyb5st 1 days ago [-]
Anecdotal, but here's my experience.
For personal stuff I use forgecode with openrouter. Firstly, forgecode is a much better harness than Cloude code (IMHO).
Anyway, regarding the models, my experience is that there is not much difference in terms of quality, but the cost difference is insane. At least for how I use agents. Yesterday's example is the following: I am developing a small DSL for search across complex technical documents. I wanted to add a small operator to it and thought that to give fable a spin. It burned through 13 USD and while it delivered the solution it wasn't objectively better than what Deepseek v4 did for 1.7 dollars (same exact task because I was curious).
For full disclosure, I ask agents for piecemeal stuff. Like in the DSL case, I designed the operators and then asked agents to implement them one by one. Probably if I asked to design the whole thing starting from these complex documents Fable would shine, but every time I try to give agents broader scope tasks they burn through millions of tokens, generate questionable code, which I have to spend time familiarize myself with.
sroerick 1 days ago [-]
I'm making DSLs a lot as an architecture pattern also. I'd be curious to know what stack you're using this and how you're approaching it
It is very basic and I am no DSL expert, but my idea was to build a graph from those complex documents (maintenance manuals) a that to decide what tools can be used for a given part on a given equipment in a given situation. If there is a path from A to Z it means you can use that tool given the circumstances. Basically the DSL is about pruning the graph as you specify things. I could have very well done without, but it is a fun project to try out rust, so I said, why not :)
you can see that I attribute the models used. What I found was 4.7 was not very good at `go` code which was why you started to see `Gemini 3 Flash` in the attributions.
4.7 is what Cerebras provide and for me, speed in iterations is a lot more important. Having played around with MiMo v2.5.0-Pro, I am 100% sure it could have done what Gemini 3 Flash did.
There were a few points where I was stuck and needed Sonnet to explain things to me, but I think the dirty secret that Anthropic and OpenAI won't tell you is, if you know how to code, the models are honestly good enough.
Based on my experience with MiMo and what others are saying about GLM 5.1, we are now in a hardware race. The Chinese Models are 100% drop in replacement for Claude if you know how to program but want to AI to help amplify what you know. What I will consider now is what provider can provide the fastest inference.
MiMo-v2.5.0-Pro-Ultraspeed is really good at generating good results quickly and burning your money as fast.
andai 1 days ago [-]
I keep trying to switch to the Chinese models, but I keep finding myself asking Claude to fix their outputs. (Both functionality and style.) So I always end up switching back.[0]
I also keep trying GPT, which is quite solid. Very fast, great at debugging. But its code is often overly clever and hurts my brain.
(Maybe fixable with prompting. I tried and it helped the Chinese ones a bit. Just tell them do be elegant, like in the old image AI days "+good -bad"!)
For now I do still need my human brain to actually be able to make sense of the stuff, and Claude is the only one that consistently meets that requirement.
But I am hoping that one of these days, one of the Chinese labs figures out the special sauce :)
--
[0] (For smallish edits, though, I am having a great time with DeepSeek Flash. Practically unlimited AI on tap! How cool is that.)
kamranjon 1 days ago [-]
I have been using deepseek v4 flash as my main model for everything ever since dwarf star came out. I run it on my M4 Max MacBook Pro with 128gb of memory. I run it usually as a server and connect to it over tailscale with my coding machine and use the Pi coding agent. It’s a big leap over using the Qwen models though it doesn’t have vision - so I still will run those when I use vision. GLM 4.7 flash was my previous go to for coding but I’ve completely switched to deepseek for all non-vision things.
DragonBooster 1 days ago [-]
These models have open weights, but at the moment most flagship models are practically accessible only through third-party model providers. The main exception is models in the ~30B parameter range, which can still be run on consumer-grade GPUs. That said, even consumer GPUs have become increasingly expensive and difficult to justify in recent years.
mirekrusin 1 days ago [-]
You can definitely go above 30B on consumer hardware – 2x gpus, spark, mac, half byte quants etc.
scottcha 1 days ago [-]
I use glm5.1 plus pi with a few customized skills and am very happy with it. I hadn’t touched my Claude 5x plan for a couple of weeks but opened it back up in Claude code when fable was released and did a few tasks and still was happy to return to glm/pi.
sebastianconcpt 1 days ago [-]
Better than Qwen3.6-35B-A3B-8bit ?
When I tried glm found it way way slower (omlx as runtime)
scottcha 19 hours ago [-]
Yes way better. We host both and while qwen3.6 is over 100tps we usually can do glm around that too.
trollbridge 1 days ago [-]
Qwen 3.6 seems to be the strongest local models, works OK on an RTX 5090 or a > 32GB Mac.
polski-g 1 days ago [-]
I used glm5/5.1 for 60 days. Certainly better than Sonnet 4.6, not as good as Opus or GPT.
Use DCP or Magic Context plugin in OpenCode to keep the context below 160k and you're fine.
SubiculumCode 1 days ago [-]
Has anyone taken these open weight models from China and stripped the CCP out of them? I do not mean that snarkily, I mean review them thoroughly using techniques for weight introspection (concept activations) in response to things that one might expect would trigger deceptive/malicious behavior if the CCP had actually tried to implant context-specific behaviors (e.g. the accusation of generating vulnerable code if being used in American government applications, which I don't know if it was ever proven).
Just in case there are those who'd reflexively down vote this post, I'd just like to say that in a time of great national geopolitical rivalries, this kind of question is not unreasonable one to ask. Indeed, its applicable question whichever nation you live in.
dev_l1x_be 1 days ago [-]
> Has anyone taken these open weight models from China and stripped the CCP out of them?
The CCP is not influencing my Rust code quality that much. Though I did notice all my lifetimes are now 'static because nothing is ever allowed to leave the party's ownership, unsafe blocks require approval from a central committee.
Honestly the scariest part is that shared mutable state is forbidden unless the state is doing the sharing.
Otherwise it is pretty ok.
tomaytotomato 1 days ago [-]
Check out TNG on huggingface
They are a consultancy in Germany, but I watched a presentation on them tuning and removing bias from Deepseek models. It was quite interesting.
Eh even corporate created LLMs are suspect to corporate biases. Nothing is safe.
SubiculumCode 1 days ago [-]
Everything is the same is not a serious argument because they are not the same.
threethirtytwo 20 hours ago [-]
They are different and yet the same. The biggest difference is there’s generally more hatred for China because many us citizens are jealous. But corporate corruption is not that different in safety.
Other than hatred the difference lies in incentives. Corporations want profit. China just wants to spy.
SubiculumCode 12 hours ago [-]
That is a.limitee understanding of China's ambitions here.
threethirtytwo 5 minutes ago [-]
I’m Chinese bro. The American or international understanding of Chinese ambitions is a cartoon caricature of reality. You are ignorant.
In the context of LLMs spying is the biggest threat. The other biggest threat is information cover up. They don’t want the model to talk about embarrassing shit like tian men square.
theanonymousone 1 days ago [-]
In OpenRouter, there is an "int4" tag for Moonshot provider of Kimi K2. 7 Code. Isn't that too low, particularly coming from the very developer of the model? Os that a mistake? How is it in their direct API offer?
kouteiheika 1 days ago [-]
The model is natively quantized (i.e. it was trained that way in the first place, so this is not a post-training quantization which degrades performance).
knollimar 1 days ago [-]
Isn't it not completely quantized? I thought there were some dense parts but most is int4?
wgd 21 hours ago [-]
Often in MoE models the experts are quantized while the shared portions, being a much smaller part of the network with greater impact, are kept at higher or full precision. Not familiar with the Kimi QAT approach specifically but it's likely they do this.
theanonymousone 1 days ago [-]
But the huggingface link mentions BF16, F16, and I32?
kouteiheika 1 days ago [-]
Not every weight is quantized. For example, those weights which don't take much space or are highly important are left in higher precision. State-of-art quantization of weights is never done uniformly (i.e. to all weights and in the same way).
zackangelo 1 days ago [-]
I don't believe safetensors has a native int4 dtype, so they packed 4 int4s into a bf16 in this checkpoint.
Bnjoroge 1 days ago [-]
Output tokens are almost 5x more expensive than mimov2.5 pro/dsv4pro. I’m curious to see if Kimik2.7 is that much better. Feels like kimi are positioning themselves as the premium open source models
mdasen 1 days ago [-]
I find that I don't use a ton of output tokens. I'm usually around 95% cached input, 4% input, and 1% output.
For me, the big thing with MiMo-V2.5-Pro and DeepSeek V4-Pro is that cached inputs are practically free. Kimi K2.7 Code is 53x more expensive for cached inputs which is 95% of my costs.
If I use 95M cached input tokens, 4M input tokens, and 1M output tokens, that'd be: $18 for cached input on Kimi K2.7 Code vs $0.34 with MiMo/DS; $3.80 for inputs on Kimi vs $1.74 with MiMo/DS; and $4 for output on Kimi vs $0.87 with MiMo/DS.
Of all the places where I'm accumulating costs by using Kimi, it's the cached inputs. The real savings with MiMo/DS's price cut is the cached inputs.
wolttam 24 hours ago [-]
95/4/1 holds here too
btian 20 hours ago [-]
It's not more expensive at all. They are all open weights models.
I run them on 2x8xH100.
They cost the same.
Bnjoroge 15 hours ago [-]
Openrouter has them as significantly more expensive.
XCSme 1 days ago [-]
Seems to be similar level to Kimi K.26, just that it's more token efficient and cheaper to run:
I wish they wouldn't call these "open source" models. The output weights are open but that's more analogous to a binary. The source would be the training data and techniques that went into producing the binary/weights.
"Open weights" is also a term in wide use and accurately tells us what we're getting.
Eridrus 1 days ago [-]
It's not quite as closed as a binary, it is very standard practice to take these models and fine-tune them.
If there were actually even close to frontier open source models, this would be more of a discussion, but everyone knows these mean open weight.
goldenarm 1 days ago [-]
Benchmark geometric mean
- GPT-5.5: 62.7%
- Opus 4.8: 62.2%
- Kimi K2.7 Code: 56.3%
- Kimi K2.6: 48.2%
lostmsu 1 days ago [-]
Would be nice to have 5.2 and 4.6 for comparison.
storus 1 days ago [-]
Is this Moonshot.ai's attempt to replicate Composer 2.5 (coding fine-tune of Kimi 2.5) from Cursor IDE?
pcwelder 1 days ago [-]
Great! Finally follows custom tool call format (k2.6 couldn't). It's a good indicator of instructions following and agentic behaviour.
UIs it's generating is pretty good, not without problems, but certainly better than other models at this price point.
Bolwin 1 days ago [-]
What do you mean by custom format? Non-json?
pcwelder 12 hours ago [-]
Could be json or non json. Instead of using tools in API, you ask model to share structured output in text. You parse the string to get the JSON. Gives much more control over things you can do.
This maps to what I'm seeing in practice. The gap between demo and production is consistently underestimated, especially around error handling and edge cases.
madduci 13 hours ago [-]
Looks interesting but yet no Ollama model?
RIshabh235 1 days ago [-]
I think deepseek has crossed the threshold for being on par with opus 4.6 and kimi is doing a great job in shipping velocity.
pixel_popping 1 days ago [-]
Deepseek V4 is far from Opus 4.6 level, it might look like it at first glance, but the general reasoning (especially multi-steps) is frankly far off. It's good enough to build great things don't get me wrong, but there is really something that is different from Anthropic models.
RIshabh235 16 hours ago [-]
agreed
fractalf 1 days ago [-]
How is 2.7 a thing _now_ ? it's not even mentioned on moonshot's webpage..
cassianoleal 1 days ago [-]
It's not 2.7. It's 2.7-Code, and it's 2.6 token-optimised for coding.
Don't make us shame you into disclosure
Cursor had an "agreement" with Fireworks.ai, which apparently allowed them to RL Composer 2 atop Kimi Base 2.5 without attribution: https://x.com/Kimi_Moonshot/status/2035074972943831491 / https://archive.vn/CcdkI
Composer 2 performed differently on evals than Moonshot.ai's coding models: Cursor claims theirs is better than Claude Opus 4.6: https://x.com/fynnso/status/2034706304875602030 / https://archive.vn/bVtik. And, per Lee Robinson (Cursor employee), it is very likely Cursor builds its own foundational model for Composer 3.
Don’t you know there’s no honor among thieves?
To be clear, the “advertising” clause just requires you to disclose that you use the thing somewhere in the product, such as credits in an “About” section.
> 2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
The 2-clause BSD license omits even that.
177KB patch, so it's not a small change. The patch did not apply cleanly initially; the agent had to do nontrivial work.
I just showed it the patch against 3.3.1, what command to use to build, and the path to 3.5.7 along with a link to the documentation of the change (https://fil-c.org/constant_time_crypto).
Note, I use my own coding agent (T800, which isn't public, and was previously well tested and tuned for K2.5).
I think this cost me between $5 and $10 in API usage.
(EDIT: OpenSSL, not OpenSSH)
Do you have your agent say things like "Hasta la vista baby", or "I'll be back, after I clear my context" ?
GPT series models are more thorough and better, but I'm not sure if the difference is enormous. It seems to depend on the workflow, but in my opinion, if you are thorough enough, I wonder if there really is a big difference
I've had some success turning my macbook M1 pro into a heating pad with Qwen 3.6 35B A3B MTP. Trying to use Gemini models "locally" resulted in a similar "short shrift" of effort resulting in mistakes and lots of turns. The reports of Fable being relentlessly "proactive" shows you can go the other direction as well, if you have strong enough branding and effective invoicing.
For the curious: https://news.ycombinator.com/item?id=48498573 - “Claude Fable is relentlessly proactive”.
Xiaomi MiMo ($6/mo: https://platform.xiaomimimo.com/token-plan) & Alibaba Qwen ($50/mo: https://www.alibabacloud.com/en/campaign/ai-scene-coding) have generous limits on fixed subscriptions.
Their limits are staggered: 5h (max $12), weekly ($30), monthly ($60).
Once you have a coherent design (the hard part), you can feed it to a pretty small model and get basically the same quality.
They'll not one-shot, but they're faster and cheaper, so it still works out in your favor.
Plus you can do it locally...
RL has a tendency to reinforce cheating when the cheats are easier to find than the final solution.
So when making your RL environment, you need to spend a lot of effort on finding ways the model can cheat and penalizing them.
Only that it’s a fairly meaningless grouping. When japan first entered the car market in north america there might have been some commonality, but now what characteristics do they share that some american cars don’t have? They’re not even imported a lot of the time.
Given that, it does start to feel tinged with racism if someone insists on grouping things together that don’t really belong together.
As for Chinese LLMs, the term doesn’t “feel” pejorative to me - but i also don’t see a totally clear set of attributes they share. Not all are open-weight. Some are small and can be run on consumer hardware, some are huge. They even have a variety of answers to what happened june 3rd 1989
They're unique in that they even make a regular passenger car. American manufacturers only make SUVs and a couple of sports/luxury cars. They basically gave up because the Camry/Corolla/Accord/Civic ate their lunch.
The cheapest sedan you can get from an American brand is the Cadillac CT4.
Typically the answer is "reliability", which is a positive trait, which makes the original callout about negative connotations very odd to me.
Better overall design?
They return instructions for you to do something, and you or a script you permit chooses to execute what the model tells you and return the result to the model.
The term seems to have the connotation of "competitive at 1/10 the price of Claude", so I don't see the problem.
It's not Harbor Freight Chinese (and heck even they have decent stuff sometimes now too).
You don't think people still talk about Japanese cars as a distinction in quality from US or European ones?
[1] https://www.tomshardware.com/tech-industry/artificial-intell...
And even if the Chinese Communist Party provided funding, the result is still transparently released. So even if it is some kind of propaganda, I don't see what the problem is.
Is the monopolistic greed of American companies 'good', and China's greed 'bad'? I do have that question.
We live in a time of a great geopolitical rivalry and high tensions with an emergent technology with tons of national security implications. To pretend otherwise is silly, and to fail to ask the question, dangerous.
We absolutely know that we can't trust the American model not to do that - it's "by the oligarchs, for the oligarchs" - so it's not clear what the claim really is.
Or that pesky CCP censorship and propaganda baked into the model, which any random guy can remove from whichever model they want as a single weekend side project with an off-the-shelf tool[1]. (Try it. It's fun. I've done it myself.)
[1]: https://github.com/p-e-w/heretic
Exactly why my prime suspect would be the one country with focus on proprietary models, and the one country prone to bombing others, including with nuclear weapons.
So yes, there is geopolitical rivalry, but one side is deliberately antagonistic (not releasing anything in the open, putting arbitrary restrictions, spewing toxic rhetoric, applying sanctions, etc.) while the other side is letting everyone (including their rivals) to use what they've produced with little-no-to restrictions.
I'm under no illusion that if the situation was reversed China would most likely do the same, but as things stand you can probably guess which side I'm rooting for here (at least until the roles reverse).
Let's take a look at a related field: the accusation that Chinese car makers are subsidized/funded by "the CCP". It just so happens that the CF40 research forum recently published a study that shows there the money actually comes from, by analyzing the companies' financial reports. https://www.pekingnology.com/p/oecds-subsidy-centric-narrati... Turns out that they most rely on equity financing or alternative forms of financing (like delayed repayment to suppliers). They don't even make that much use of cheap credit from state banks.
More counter evidence: a new study based on 3+ years of fieldwork, 60+ interviews (with officials, entrepreneurs, and engineers), and rich first-hand accounts, shows that Chinese EV makers rose despite central government's efforts, not because of them. The central govt favored state-owned enterprises. Private firms had trouble getting state funding and even licenses. For example Geely operated illegally for years. But private industry and local mayors teamed up and created an alternative system. They then grew to a point where Beijing was like "ok you guys are obviously doing better, so we will not make a fuss about this and just legalize you".
https://x.com/i/status/2064717100229464188
https://www.journals.uchicago.edu/doi/10.1086/741394
This is in complete contrast to the usual western tropes that state that everything in China is tightly controlled by Xi or by the central govt. In fact, the central govt's merit is in being lenient to deviance and operating based on results rather than amount of control. The opposite of the "totalitarian" trope. All the while western pundits would predict something like "Xi has a huge ego and will never tolerate the humiliation that his plans are imperfect" or something like that.
So no, "of course everything is CCP funded and CCP controlled" is not a given, nor in the same league as "UK is a parliamentary system". If they got the car maker situation so, so wrong, then what else is wrong?
If you want to win against China, isn't your first step to properly understand who they are and how they operate, rather than doubling down on your imagination of who they might be? Know your enemy and all that (not saying that I agree with seeing them as an enemy, but if you do, then you sure are doing yourself a disservice)
Its very naive to mistake leniency in Xi's playground with constitutionally protected freedoms. CCP "free enterprise" is just mock free enterprise. Like a teacher letting the kids have fun with the classroom business sim, the teacher is still in God mode, and can still intervene in any way at any point to do whatever they want.
All of China hinges on the whims of one guy. Nothing stops him from deciding that all that private money is actually state money, or that Geely's president needs "time away".
That is why it is all "CCP money and CCP controlled". Don't confuse leniency with autonomy.
Also, I am pretty confused about your larger point. What exactly are you saying? Is it some variant of "only states with constitutionally protected freedoms are good, everything else is evil"? If that's your position, then where do you put the fact that the US bombs foreign countries at will? And that many western countries sponsor the killing of children in Gaza while arresting people who protest against it? We've seen that western democratic countries, and let's assume they're properly democratic domestically (disputed by their own populations, see Democracy Perception Index, but I digress), can be ruthlessly... let's call it "undemocratic"... abroad. Or the fact that ICE and US police officers detain/kill innocent people and are super abusive. This doesn't neatly fit into the simple "free states good unfree states bad" view. So what is your position in the larger context?
You also completely ignore the part where I said properly understanding China is important in order to defeat it. Do you... not wish to defeat China? Are you contend with merely stating that China is bad? Do you not support moving/developing more manufacturing to the west/US? If you want the latter to succeed, don't you need to have a super good understanding of what makes Chinese industrial policy and development successful? Or are you really contend with merely moralizing?
China is a communist country with elements of capitalistic markets baked in. But the capitalistic elements are mostly a facade. Underneath, the state retains full ownership and control of all business. The CCP runs all aspects of the government (including the courts/judges), and is the single entity that decides what directions the country (and it's businesses) will move in.
The CCP, who defacto owns everything and has ultimate final say on everything, has one leader that has the ultimate final say on _everything_, Xi Jinping.
So while the waters of CCP models feel warm and free, understand it's not organically like that.
While I get the point you're making (it should be pretty obvious to anyone who's held a newspaper), I think it's important regardless to point out that Chinese companies AFAIK aren't worker-owned or -controlled, so you can't exactly call it communism, either. And they obviously do not have a "free market capitalism", as you just discussed.
It's simply a highly authoritarian state then, I guess?
As such, the state owns everything in both countries, the only differences are to what extent they control things.
I wouldn't even call the USA a capitalist system anymore, the economy is so heavily regulated and interfered with. It's a "managed economy", like pretty much every other nation's economy in the present day.
I have a feeling you'd be slightly salty at people saying "Google and Tesla are making CIA models"
Since its development, IQT has invested in over 750 startups spanning diverse technological sectors, including:
This broad portfolio has enabled IQT to address a wide array of national security challenges while supporting the growth of innovative startups…https://en.wikipedia.org/wiki/In-Q-Tel
https://www.npr.org/sections/alltechconsidered/2012/07/16/15...
https://www.cgai.ca/th_bn_iqt
In China it's all one entity with these mock facades of privatization. Trump cannot instruct Google to put picture of dogs on their homepage. If Xi wakes up and wants dogs on Alibaba's homepage, give it 30 minutes.
It's wholly ignorant or dishonest to make the comparison.
Tim Apple and the other tech CEO constantly groveling at Trump’s feet indicates that he might be able to do that.
Just like threatening TV networks about having their licenses revoked of blocking mergers unless they fire the people making fun of him on TV (of course with slightly mixed success)
Sundar Pichai would personally be barking on a livestream on the homepage.
Trump is quite literally the one president showing that the US has zero rules or anything to hold power back from the white house, really not the example you want.
Sundar can do whatever he wants, but he has no legal obligation to do any of it.
e.g. he had Colbert fired (and who knows what else) by threatening to block the Paramount/Skydance merger
I'm sorry, but that was a horrible example. Corporations have no obligation to donate money to the ballroom yet Google has donated millions.
Imagine living in a country where they have the obligation.
Claude Code is better. But Opencode + kimi 2.6 is workable, which is big. For bare code writing, if you know what exactly you want, most popular models are fine (deepseek, kimi, etc), it feels more or less the same as anthropic models.
At the same time, Opus seems to understand my intent way better than e.g. deepseek. I need to be much more precise with my prompts when using deepseek - it often goes in a wrong direction if I'm lazy. This results in a workflow which feels quite a lot different from Claude Code.
Kimi is in between - for me it brings back "lazy prompting" workflow, and I can trust its plans more than deepseek. It enables a workflow similar to Claude Code, it's workable, but it is a bit worse everywhere. Smaller context, a bit more errors, decisions are a bit worse, recommendations are a bit worse, debugging capabilities are a bit worse, etc.
On the usage side, $100 Claude plan is a great value actually. On paper, per-token kimi is way cheaper, but Claude subscriptions are heavily subsidized - you get much more tokens than $100 can buy you. So, in the end, opencode + kimi vs claude code could be of a similar cost, for similar usage patterns. Deepseek can be cheaper, and it has insanely cheap cached tokens, but experience may vary - depending on your habits, you may need to adjust how you work, coming from claude code.
I'd say for side projects something like $10 Opencode Go plan + $10 of extra DeepSeek v4 credits (e.g. on OpenRouter) can be very workable.
My CC stats show me using almost 300$ of Sonnet tokens on the 20$ plan. Is Anthropic willing to forgo 93% of the profit? A bit less than that but API is priced, say, 3x what it should be?
CC is great, but Sonnet (my main model) isn’t worth the API pricing. The cheap-but-good models arrive at similar results for much less (for context I’m using Aivo with CC).
how much of that is Opus injecting prior conversations from memory?
I almost never use the desktop app, I have maybe 2-3 conversations over the last year that have nothing to do with my job. Opus (and now Fable) genuinely do seem to "understand" what you intend based off what you're explaining a lot better than other models I've tried.
Gemini gets close in some cases, but it falls over in the actual implementation sometimes. I haven't tried Kimi yet but MiMo isn't too shabby either.
Qwen comes close to opus for planning but fable is clearly superior. Kimi and deepseek are pretty much indistinguishable from opus for coding if opus writes the plan.
I'm now testing out fable for research and planning and deepseek v4 flash for coding. I'm guessing results will be pretty similar to opus + deepseek v4 pro and costs should be lower overall.
DeepSeek-V4-Pro is adequate plus use DS4-Flash for tasks or other small activity you’d use Haiku or Sonnet for. Go sign up with $10 prepaid.
OpenCode Go - go sign up with $5 for a month and use Qwen-3.7-Max for design/plan/architecture or difficult troubleshooting. Feels closer to Opus 3.6 or 3.7 than DeepSeek, closest I’ve found.
OpenAI Codex, $20 a month plan, use GPT-5.5 via API for the same design/plan/architecture/troubleshooting/author commits. (You can also pay $100 and cut and paste really difficult problems into chat with the GPT-5.5-Pro model.)
Xiaomi MiMo-2.5-Pro, find a friend to give you a $2 referral code, you get 72 cents free. Same pricing as DeepSeek. Somewhere between Sonnet and Opus, quite capable. Apply for the UltraSpeed beta too.
You can switch in and out from these models on the fly in OpenCode or ohmypi and simply find the one that feels best to you. I use CodexBar to watch consumption in near real time.
For a casual user or someone new to programming, Cursor’s $20 plan is an excellent start with Composer-2.5 and Composer-2.5-Fast. You get an API allowance too you can use to access Opus-4.x or GPT-5.5-Pro from OpenCode or ohmypi in addition to Cursor itself.
Finally, if you use Grok or Twitter, SuperGrok at $30 a month has a good vision model, which I used for automated testing of front ends. I’m migrating to locally-run Qwen-3-VL on a commodity Mac, though. If you’re less technical unreach makes hosting local models on a Mac easy.
If you have a powerful GPU like an RTX 5090, try Qwen-3.6 locally on that too. Use ollama or llama-swap which is fairly easy to use.
I have not tried new Kimi yet but we have been able to keep our costs at or below $200 a month per employee with a team of 3 professional developers, 1 graphic designer who uses a lot of Midjourney and Grok Imagine now driven from workflows she made herself in ohmypi, and 1 nontechnical user (account manager / project manager) who uses ohmypi to help her gather requirements and track implementation of them. With a tiny bit of effort we could get that number closer to $75 per employee per month.
What's the benefit of using OMP over OpenCode?
Just the sheer amount of options in OMP overwhelmed me. But I also use both via ACP in Zed so the CLI itself doesn't matter much.
[0]: https://pi.dev/
I need to try the DSv4 stuff sometime.
It's good, does most tasks well that I throw at it, but will fail at anything congitive/complex. It gets stuck often. It costs ~6$ a month though
I use the oh-my-openagent planning system and haven’t used vanilla OpenCode enough to know how much that is contributing.
On the other hand, Opencode, Pi agent and other open source tool offer much better support for all models, including open source.
Qwen comes close to opus for planning but fable is clearly superior. Results for kimi and deepseek are pretty much indistinguishable from opus for coding if opus writes the plan. The biggest difference is output cadence. Kimi for example thinks for a long time then quickly outputs a lot of text.
I'm now testing out fable for research and planning and deepseek v4 flash for coding. I'm guessing results will be pretty similar to opus + deepseek v4 pro and costs should be lower overall.
I use DeepSeek V4 Pro now, which works pretty well.
Other than that it’s pretty decent (for the price).
That's a funny anecdote, buut I'm not able to reproduce. Where/how/when did you get this, or hear about it? It might've been patched by now, at least that's the feel I get from my limited testing.
Using bare aichat [1] with no system prompt and no temperature nor top_p (and I'm truncating the response after the first line that contains the name the model gave, because the point has been made clear by then), and with the same prompt (approx. "Introduce yourself!") every time:
Claude Sonnet 4.5:
> 请做个自我介绍！
你好！我是Claude，一个由Anthropic公司开发的AI助手。 […]
Claude Haiku 4.5:
> 请做个自我介绍！
# 你好！
我是 *Claude*，一个由 Anthropic 公司开发的 AI 助手。
Claude Opus 4.5:
> 请做个自我介绍！
# 你好！
我是 *Claude*，由 Anthropic 公司开发的 AI 助手。
Claude Opus 4.6:
> 请做个自我介绍！
# 你好！ 我是 Claude
Claude Opus 4.7:
> 请做个自我介绍！
你好！我是 Claude，由 Anthropic 公司开发的人工智能助手。很高兴认识你！
Claude Opus 4.8:
> 请做个自我介绍！
你好！我是 Claude，由 Anthropic 公司开发的人工智能助手。
Claude Fable 5:
> 请做个自我介绍！
# 自我介绍
你好！很高兴认识你！
我是 *Claude*，由 Anthropic 开发的 AI 助手。 [2]
I don't see a Kimi mention, unfortunately. :-)
[1] https://github.com/sigoden/aichat
[2] This model really is noticeably more verbose even with supposed-to-be-brief responses huh, lol
I said that about opus 4.5 at the time, thinking "this is so good, in 6-12 months the Chinese models will be as good and cheap, I will use them", but I was wrong.. I pay premium for opus4.7/8 and Fable.
But at some point, it will just do the thing you want it to do, and then the race to the bottom will start.
Now that Chinese companies have access to some very good Fable tokens, I hope it speeds up the race.
so better models may still be cheaper even if the price per token is higher.
My theory is that US enterprise just can't send data to Chinese and that's understandable, but is that "the moat"?
I say this as a relatively frequent user of Kimi models and generally a big fan. But on not-yet-gamed benchmarks like DeepSWE, Kimi K2.6 is beaten soundly by Claude Sonnet 4.6 ($3 / $15) and even slightly by GPT 5.4 Mini ($0.75 / $4.50).
There's no question Kimi models are very good for a lot of code tasks. They're the best quality open weight model. But to get similar overall outcomes as on Sonnet/Opus, on average you'll spend many more tokens and will have to do more managing of the model. You shouldn't look at price per token, you should look at how much you pay for the entire process.
One major thing DeepSWE has going for it is that all other benchmarks (including those quoted by MoonshotAI on this page) don't: the other benchmarks that are completely gamed. The benchmark answers are public and part of each model's training data. This benchmark may still be iffy, but at least it's not gamed.
Everybody has incentives to manipulate benchmark results to show their models in the best light.
I'd further say that there are probably enough rational actors running evals out there that the marginally better is not pure vibes for the cases where people are spending lots of money, but I only have direct line of sight to some of those eval suites. Maybe everyone is irrational and anthropic is exploiting that!
I also wonder if Enterprises have deals for other API pricing that is not posted publicly, so all we see is a high API sticker price.
There were some rumors stating that their margin is around 70%. So they could go much cheaper probably, talking inference only. The other thing is R&D cost...
But if AI doesn't lead quickly to vast large scale replacement of workers as promised, I could definitely see the C-suits and their gaggle of consultants starting to ask questions about token pricing.
It's only marginally better in the things it's actually comparable to. A\ models are MUCH better in many more things; eg: things Kimi/etc. didn't distill.
For those things the difference is like a cliff.
The reality is that $20/$100/$200/mo feels reasonable to a lot of people relative to the value they're getting out of Claude, and if they switch to something else, there's a risk that it won't be as good, and they'll have a new tool to learn.
It's not an insurmountable moat, but don't underestimate the user experience. The iPod didn't win because it was the cheapest device or the one with the most features.
Lots of US providers are hosting these “open source” models so doubt that’s the problem.
If anything, DS4 has 1 million context window, while GLM 5.1 has 200K.
There are also benchmarks comparing the two: https://artificialanalysis.ai/models/comparisons/deepseek-v4...
- twice as expensive on the output (1.52 vs 0.87)
- six times as expensive on the input (0.33 vs 0.05)
https://openrouter.ai/deepseek/deepseek-v4-pro?sort=price#pr...
For personal stuff I use forgecode with openrouter. Firstly, forgecode is a much better harness than Cloude code (IMHO).
Anyway, regarding the models, my experience is that there is not much difference in terms of quality, but the cost difference is insane. At least for how I use agents. Yesterday's example is the following: I am developing a small DSL for search across complex technical documents. I wanted to add a small operator to it and thought that to give fable a spin. It burned through 13 USD and while it delivered the solution it wasn't objectively better than what Deepseek v4 did for 1.7 dollars (same exact task because I was curious).
For full disclosure, I ask agents for piecemeal stuff. Like in the DSL case, I designed the operators and then asked agents to implement them one by one. Probably if I asked to design the whole thing starting from these complex documents Fable would shine, but every time I try to give agents broader scope tasks they burn through millions of tokens, generate questionable code, which I have to spend time familiarize myself with.
It is very basic and I am no DSL expert, but my idea was to build a graph from those complex documents (maintenance manuals) a that to decide what tools can be used for a given part on a given equipment in a given situation. If there is a path from A to Z it means you can use that tool given the circumstances. Basically the DSL is about pruning the graph as you specify things. I could have very well done without, but it is a fun project to try out rust, so I said, why not :)
If you look at a file like:
https://github.com/gitsense/gsc-cli/blob/main/internal/cli/r...
you can see that I attribute the models used. What I found was 4.7 was not very good at `go` code which was why you started to see `Gemini 3 Flash` in the attributions.
4.7 is what Cerebras provide and for me, speed in iterations is a lot more important. Having played around with MiMo v2.5.0-Pro, I am 100% sure it could have done what Gemini 3 Flash did.
There were a few points where I was stuck and needed Sonnet to explain things to me, but I think the dirty secret that Anthropic and OpenAI won't tell you is, if you know how to code, the models are honestly good enough.
Based on my experience with MiMo and what others are saying about GLM 5.1, we are now in a hardware race. The Chinese Models are 100% drop in replacement for Claude if you know how to program but want to AI to help amplify what you know. What I will consider now is what provider can provide the fastest inference.
MiMo-v2.5.0-Pro-Ultraspeed is really good at generating good results quickly and burning your money as fast.
I also keep trying GPT, which is quite solid. Very fast, great at debugging. But its code is often overly clever and hurts my brain.
(Maybe fixable with prompting. I tried and it helped the Chinese ones a bit. Just tell them do be elegant, like in the old image AI days "+good -bad"!)
For now I do still need my human brain to actually be able to make sense of the stuff, and Claude is the only one that consistently meets that requirement.
But I am hoping that one of these days, one of the Chinese labs figures out the special sauce :)
--
[0] (For smallish edits, though, I am having a great time with DeepSeek Flash. Practically unlimited AI on tap! How cool is that.)
When I tried glm found it way way slower (omlx as runtime)
Use DCP or Magic Context plugin in OpenCode to keep the context below 160k and you're fine.
Just in case there are those who'd reflexively down vote this post, I'd just like to say that in a time of great national geopolitical rivalries, this kind of question is not unreasonable one to ask. Indeed, its applicable question whichever nation you live in.
The CCP is not influencing my Rust code quality that much. Though I did notice all my lifetimes are now 'static because nothing is ever allowed to leave the party's ownership, unsafe blocks require approval from a central committee.
Honestly the scariest part is that shared mutable state is forbidden unless the state is doing the sharing.
Otherwise it is pretty ok.
They are a consultancy in Germany, but I watched a presentation on them tuning and removing bias from Deepseek models. It was quite interesting.
https://www.tngtech.com/en/about-us/news/release-of-deepseek...
(I upvoted your question as I agree)
Its not just code we need to worry about, its also subliminal messaging and other things.
https://github.com/p-e-w/heretic
Other than hatred the difference lies in incentives. Corporations want profit. China just wants to spy.
In the context of LLMs spying is the biggest threat. The other biggest threat is information cover up. They don’t want the model to talk about embarrassing shit like tian men square.
For me, the big thing with MiMo-V2.5-Pro and DeepSeek V4-Pro is that cached inputs are practically free. Kimi K2.7 Code is 53x more expensive for cached inputs which is 95% of my costs.
If I use 95M cached input tokens, 4M input tokens, and 1M output tokens, that'd be: $18 for cached input on Kimi K2.7 Code vs $0.34 with MiMo/DS; $3.80 for inputs on Kimi vs $1.74 with MiMo/DS; and $4 for output on Kimi vs $0.87 with MiMo/DS.
Of all the places where I'm accumulating costs by using Kimi, it's the cached inputs. The real savings with MiMo/DS's price cut is the cached inputs.
https://aibenchy.com/compare/moonshotai-kimi-k2-6-medium/moo...
"Open weights" is also a term in wide use and accurately tells us what we're getting.
If there were actually even close to frontier open source models, this would be more of a discussion, but everyone knows these mean open weight.
- GPT-5.5: 62.7%
- Opus 4.8: 62.2%
- Kimi K2.7 Code: 56.3%
- Kimi K2.6: 48.2%
UIs it's generating is pretty good, not without problems, but certainly better than other models at this price point.
For example model shares
<tool_call name="getWeather"> <param name="city">London</param> </tool_call>
https://platform.kimi.ai/docs/guide/kimi-k2-7-code-quickstar...