Unfortunately I can't find sources other than this for now but this seems to be legit.
mohsen1 · 2026-08-26 11:08:16 UTC
> The company on Wednesday confirmed speculation that the Ox Alpha model is a new iteration of its GLM series and said it will release the weights for it tonight, in response to queries by Bloomberg News.
Seems legit.
It's really hard to know how good it is. So much hype around it.
eli · 2026-08-26 13:33:01 UTC
I mean, you can try it for free.
KaseyKim · 2026-08-26 11:21:39 UTC
they have confirmed it officially
dgellow · 2026-08-26 10:36:55 UTC
Do we know the size of the model?
j_maffe · 2026-08-26 10:41:06 UTC
Anyone has a link to a report of its capabilities? I can't find a reliable source.
vblanco · 2026-08-26 10:49:44 UTC
completely vibes based, but ive been using it to port Mindustry game from Java to C# with agents, and its been working for 50 hours (its 15-20 tks so super slow inference). Its done a fantastic work and its almost finished now. Better results than deepseek flash and gpt luna by a mile on this kind of long term work. Less good than gpt sol or opus. We dont know the param count but my guess is 200-300 range.
le-mark · 2026-08-26 11:58:47 UTC
Just curious, what is the motivation for this conversion?
vblanco · 2026-08-26 12:21:01 UTC
Its free tokens so i left it running for fun as a experiment
daveyoung · 2026-08-26 10:56:48 UTC
likely a distilled glm 5.3 that will punch within 20% of that at 2-3x less size. you'll find that capability is typically very jagged on models that are distilled
Wenghi is behind DeepSWE, one of the best benchmarks.
esskay · 2026-08-26 10:49:22 UTC
I'd be interested to know what was going on with it during the public test as there were numerous reports of it improving considerably at tasks it was asked to do early on in the test compared to later in it.
rfoo · 2026-08-26 10:53:36 UTC
lol don't shout out the obvious
daveyoung · 2026-08-26 10:54:23 UTC
Two potentials from my pov:
1. Just variance in pass@K. If you prompt any model multiple times you'll see a large variance. N=1, but I find chinese open source models have a higher variance than higher-RL'd models like fable/opus.
2. They legitimately shipped a new RL checkpoint over the 7 days, which I find hard to believe.
I am leaning towards 1.
re-thc · 2026-08-26 11:14:17 UTC
2. There was a new checkpoint. Official.
daveyoung · 2026-08-26 11:29:18 UTC
do you have reference to where it was said?
zarzavat · 2026-08-26 11:14:30 UTC
3. Deployment problems unrelated to the weights causing degraded performance
swiftcoder · 2026-08-26 12:47:16 UTC
For sure the version accessible from OpenCode had a massive timeout problem the first day or so, which seemed to heavily degrade its task completion rate
dannyw · 2026-08-26 13:17:07 UTC
Or 3, they find some bug/regression in their pipeline; maybe they didn't quant parts of a model properly, maybe their inference engine had a bug, maybe some pinned MoE expert wasn't pinned, etc...
That's very plausible to have, identify, and fix in a day; especially when you get community feedback in the wild.
utilize1808 · 2026-08-26 11:23:31 UTC
It's logical to serve the best version (quant) of the model at the beginning so that users keep testing it. It is also reasonable to think that the developer of the model tried to test various quant levels by gradually degrading the model's capabilities.
brookst · 2026-08-26 11:58:40 UTC
I mean that’s imaginative but not sure there’s any evidence at all for it, and it’s the opposite of what the comment you replied to observed.
WithinReason · 2026-08-26 10:52:31 UTC
Mixed signals, here it's performing below even GPT-5.4 Nano:
but if the latter is true, will people still say it was "distilled" from Fable?
sunbum · 2026-08-26 10:56:30 UTC
the 2nd website is not official, just something someone slopped together for some reason.
Alifatisk · 2026-08-26 11:11:12 UTC
I have plenty of these websites, I can’t understand why someone is doing this.
colesantiago · 2026-08-26 11:35:36 UTC
It is called phishing and grifting.
Many people and even software engineers fall for this all the time.
Most of these people are from crypto pivoting to AI doing this.
AI has made this easier and cheaper and it is going to get a LOT worse.
Imagine lots of websites with typosquatting and looking exactly the same as another website, vibe coded and cloned within seconds.
The public have no chance.
Alifatisk · 2026-08-26 16:36:50 UTC
What is there to phish? These are simple vibe coded websites providing information for a certain topic, nothing else. In this case, that 2nd url is a website with information regarding the new model as well as a broken chat interface to try out.
colesantiago · 2026-08-26 20:55:20 UTC
You do realise there are hundreds of these types of 'sites'.
This one that is listed is designed to rank on Google as an informational source (although unofficial and not from z.ai which is why I said it is phishing)
Assuming you are technical you are able to discern this, imagine the average person.
No chance.
Alifatisk · 2026-08-27 18:00:03 UTC
> You do realise there are hundreds of these types of 'sites'.
Yes, and its these sorts of websites I am asking about.
> This one that is listed is designed to rank on Google as an informational source (although unofficial and not from z.ai which is why I said it is phishing)
Again, what is there to phish?
MrDrMcCoy · 2026-08-27 22:15:21 UTC
Phishing implies exploitable data collection. Is that happening here?
yorwba · 2026-08-26 11:26:16 UTC
Even if it weren't slopped together, 65% vs 80% on 10 tasks just isn't a significant difference. For 80% power to distinguish at a significance level of 0.05, you'd need more like 140 samples, if those were the true success probabilities.
The number one problem in LLM benchmarking is that people try to draw conclusions from sample sizes far too small to conclude anything but "it works sometimes, it fails sometimes, hard to say which is better." (The number two problem is that people run benchmarks blindly without checking that they measure something meaningful.)
brotchie · 2026-08-26 18:33:50 UTC
I'm going steal "slopped together", great quip.
epolanski · 2026-08-26 11:02:27 UTC
GLM 5.3 was a great model, so this would be strange to release a regressed model
ImprobableTruth · 2026-08-26 11:06:54 UTC
It's probably GLM 5.3 flash, so weaker but cheaper.
re-thc · 2026-08-26 11:13:43 UTC
With vision on top
re-thc · 2026-08-26 11:13:15 UTC
the outperform Fable was a mid (not completed) benchmark run. Real results were lower.
woadwarrior01 · 2026-08-26 11:27:15 UTC
That benchmark is super sus. Until someone pointed it out, the top performing open weights model was a Kimi K3 fine tune from their sponsor (abacusai/Smaug-Agentic). Now, it's not on the list.
I really want to see hard evidence of distillation before I buy into it. Seems like a lot of sour grapes over not having the sort of lead assumed. In this field, it has been shown repeatedly that leaps in performance come swiftly and without notice.
xienze · 2026-08-26 11:53:22 UTC
What would constitute evidence in your opinion?
anon373839 · 2026-08-26 12:31:15 UTC
How about proof that black-box distillation can deliver these results without a very sophisticated RL pipeline doing the heavy lifting?
dannyw · 2026-08-26 12:57:39 UTC
"Black-Box On-Policy Distillation of Large Language Models", Microsoft Research, https://aka.ms/GAD-project
> 'GAD consistently surpasses standard sequence-level distillation, delivering superior generalization and achieving performance that rivals the proprietary teacher. These results validate GAD as an effective and robust
solution for black-box LLM distillation.'
No RL, although I'm a little bit surprised to see MS Research publishing a paper on distilling GPT5?
anon373839 · 2026-08-26 14:10:02 UTC
That's an interesting paper, but there is virtually no discussion of reasoning behaviors or optimization for long-horizon tasks (i.e., all of the recent advances in LLMs that people care about). The evaluation methodology also is pretty dated:
> We reserve 500 samples of LMSYS-Chat-1M-Clean as the primary test set. We also include test datasets consisting of a 500-sample subset split from Dolly [6], the 252-sample SelfInst dataset [37], and the 80-question Vicuna benchmark [3] to evaluate out-of-distribution generalization. We report the GPT-4o evaluation scores [45, 10], where GPT-4o first generates reference answers and then scores the output of the student model against them. We also conduct human evaluations on the LMSYS-Chat-1M-Clean test set for qualitative assessment.
hypfer · 2026-08-26 12:07:31 UTC
FWIW, the way GLM-5.2 (and 5.3) talk is clearly claude, so it is for sure also trained using distillation.
The metric used there is me screaming at my screen per operating hours.
Does it matter? IMO not really. Weights are open after all. (Or.. soon at least for 5.3)
dannyw · 2026-08-26 12:50:23 UTC
With the amount of Claudish on the internet now, and in source code repositories (how many Claudish README.mds have you seen?), you don't have to make a single API call to end up with a model that talks like Claude.
And critically, like contracts in general, Anthropic's terms of service is only binding upon the user/counterparty. So even if a company say specifically sought out 'claude-like' content, and claude code traces available on the internet, if they don't use the Anthropic platform there is no ToS claim.
RataNova · 2026-08-26 13:54:29 UTC
Half of the new open-source stuff on github is written by claude now, all the way from issues to docs. Models are just vacuuming up this dataset during pretraining, naturally picking up the tone. You don't even need direct distillation via api anymore when the whole internet has turned into one big snapshot of Anthropic's weights
Aurornis · 2026-08-26 12:16:26 UTC
Claims about Ox Alpha performing at Fable level were from the social media hype cycle. Everything new in the LLM space brings a wave of influencers hyping it up as a revolutionary leap forward. Don’t forget to like and subscribe to learn more.
It is a capable small model, but it’s not frontier level. The interesting part will be seeing the model size, how it responds to quantization, and how fast it runs on the kind of non-server hardware that we can buy without selling a kidney.
worldsavior · 2026-08-26 12:43:57 UTC
This influencers are getting paid, it's not coincidential.
nixon_why69 · 2026-08-26 13:41:25 UTC
They don't have to be getting paid. The natural bias of media is towards laziness and sensationalism (stolen from Jon Stewart, so maybe the same is true about comments).
kkukshtel · 2026-08-26 14:22:46 UTC
I think so much of this is people greenfield-ing things as benchmarks, which is nearly always a success case for any AI these days.
plumeria · 2026-08-26 14:37:58 UTC
> Claims about Ox Alpha performing at Fable level were from the social media hype cycle.
Kinda useless to compare simply based on model without considering harness. Different agents handle the context etc completely differently. I would like to start seeing these model vs model comparisons across different harnesses.
daralthus · 2026-08-26 12:53:25 UTC
omp+0x-alpha beat both cc+fable and codex-sol in creating/refactoring a big eval setup. the former just knows where things should belong and completed the task all the way while the other two failed on both metrics.
I've been having oxa and sol do architecture design then compare notes. Sol is definitely still way ahead. But there's reliably some really good wins ideas and concepts that OxA throws out there that Sol is very happy to encorporate.
One thing that I think matters a lot for the non developers, all three of us (sol, oxa, and me) usually agree that oxa's write up is far far better. It explains the situation very well, and has great structure for its write ups. Sol gets the job done, but it's terrible at re-explaining the problem for humans, at laying out information. It also doesn't show it's thinking, so it's imo a terrible peer to work with!
tosh · 2026-08-26 10:57:58 UTC
my guess is this is a small model punching way above its weight
on toy benches it made quite a few mistakes but was able to fix all of them on its own
(meaning more tokens, more turns, more tool calls — but same outcome as gpt 5.6 sol)
I feel it is the only model which has some humanity/empathy in many situation compared to latest "intelligent" model.
owebmaster · 2026-08-26 16:07:18 UTC
I'm risking having less "humanity/empathy" than ChatGPT 4o now but that's a socializing problem of the people feeling that way. Models are calculators, not therapists/friends.
Leynos · 2026-08-26 17:57:55 UTC
The latest snapshot of 5.6 Sol feels disturbingly 4o like at times on high in ChatGPT. Although it swears like a sailor
Xi has made it official policy, see his keynote speech at their World AI Conference last month:
> We should seize this rare, historic opportunity to encourage open source, openness, collaboration and sharing. [1]
People have pointed that this seemingly made Alibaba/Qwen turn around from closing their models (this was rumored after the shakeup early this year [2]) and release the weights for even the Max variant of their new models, which they previously did not.
Alibaba execs probably had PTSD from last time they attempted to defy CCP policy...
mschuster91 · 2026-08-26 20:40:57 UTC
Well, China is at their usual barely legal (at least if the WTO would be worth even a bit after it got thoroughly gutted) game, just burning their effectively infinite cash reserves to undercut Western providers and eventually force them out of business.
In the end there's barely any moat that any of the ludicrously "valued" AI companies have - the only thing justifying the valuations of SpaceX/xAI/Tesla/Anthropic/OpenAI is a supposed "secret sauce" that, frankly, barely exists any more.
Everyone and their dog can go and run LLMs for dirt cheap on their own hardware.
respectattentio · 2026-08-26 11:26:22 UTC
it's for sure better than deepseek flash 07/31
mark_l_watson · 2026-08-26 12:09:06 UTC
That is saying a lot if Ox Alpha is also small and relatively cheap computationally. I hope so; I love deepseek-v4-flash-0731 and use it frequently. Fast inference is good and fits with my dev style: I like to be in the loop, not let an agent code on its own for long periods of time.
SyneRyder · 2026-08-26 16:03:13 UTC
From their blog post, it's 320B total parameters and 18B active parameters, so a similar size, but slightly bigger.
Regular pricing is $0.15 input, $0.50 output... but currently 50% off, making it $0.075 input and $0.25 output. That beats most of the V4 Flash providers, but not all, and obviously tokens per task may not be equivalent.
I've also just noticed the blog post reveals the Artificial Analysis score - it's a 57, so it's Opus 4.8 / 5.6 Terra level.
It is not going to be cheaper though. I may choose the cheaper one in the end because performance will be marginal, both being flash.
harlan_pdx · 2026-08-26 11:31:56 UTC
Releasing weights is the right move. Keeps them competitive with DeepSeek on the open side.
stanac · 2026-08-26 11:47:00 UTC
I had good experience with GLM 5.3, but...
Z.AI is the only provider for GLM 5.3 on OpenRouter. I don't see 5.3 on Hugging Face. Not sure if this new model is "full GLM" or something smaller, or if they will like Moonshot AI publish weights but put restrictive license [1], which will again leave Z.AI as single GLM model provider on OpenRouter.
The release date is supposed to be August 28th 2026
cute_boi · 2026-08-26 14:09:03 UTC
I am happy if they publish under restrictive license. Developing model costs a tons of money and company need to make money somehow by still open sourcing project.
glimshe · 2026-08-26 11:36:23 UTC
There's a lot of brand confusion among the Chinese models right now. Kimi, Qwen, GLM, Z.ai, Ox. We might know the difference (or I should say, someone does because I'm losing track already) but these models have no chance at end user penetration and loyalty until there's a single focused survivor.
It took me a year talking about it until my wife knew that ChatGPT and Gemini are two different things.
PS: some replies, especially if you do a deep dive on comment history, clearly expose the joint effort to drum up support for Chinese models. This has been clear on HN lately as anything even slightly critical of Chinese tech gets downvoted unnaturally quickly. One can just wonder what's behind the effort...
giwook · 2026-08-26 11:39:44 UTC
I disagree. I think most users who are savvy enough to be using openweight models and/or running models locally are not dealing with the same level of confusion you are.
Ox is just GLM. And z.ai is the maker of GLM.
The main players in the openweight model market have been known for a while.
And they already have significant user penetration.
tokai · 2026-08-26 11:45:44 UTC
Just because you're confused doesn't mean that there is general confusion here. Its really not that complicated.
marclove · 2026-08-26 11:46:15 UTC
Consumers aren’t the customer.
mark_l_watson · 2026-08-26 12:05:50 UTC
I have seen studies from MIT and Stanford that the majority or US startups are using much less expensive open weight models so consumers of their products are open model users whether they know it or not. These are often Chinese models.
Not to go off topic but I am pleased to see open model support from US companies like Poolside.ai, NVIDIA, IBM, Google, etc.
seaal · 2026-08-26 12:11:24 UTC
There's a lot of brand confusion among the American models right now. ChatGPT, Claude, Gemma, OpenAI, Meta, Google, Muse Spark, Anthropic, Microsoft, Gemini. We might know the difference (or I should say, someone does because I'm losing track already) but these models have no chance at end user penetration and loyalty until there's a single focused survivor.
It took me a year talking about it until my wife knew that Kimi K3 and GLM 5.3 are two different things.
giwook · 2026-08-26 12:21:35 UTC
Well done.
Terretta · 2026-08-26 12:11:35 UTC
The bubbling froth at the open edge is getting user adopted at a crazy pace, by the early adopter persona trying them all within hours to days. This persona loves taking apart and putting together novel things, and telling others.
Fast follower persona clusters around emerging zeitgeist across the tellings. At the moment, arguably that's mostly Qwen for everyday hobbyists, and GLM for those that can run 512GB to 1.5TB of memory. This persona is seeking viable applied results: "I have frontier at home".
The early majority pick things up after models are curated into apps like LM Studio or one's platform app of choice, usually at least one major release behind because it takes that long to choose and package into mass distribution.
This is the step where early majority persona "has no idea" what the parade of weird names is about, they care about qualia of the conversations they try to have.
This persona is, at present, very under-served, and likely to remain so until mass devices can perform feeling like 27B at Q4 large quality better, or workplace devices can achieve a pragmatic utility like 135B at Q8 or better.
Harnesses that work where the workplace persona lives bridge this. This persona doesn't care the Chinese model name, they care "does it code?" For that, the applied harness and model take time to be matched, as JetBrains did harnessing a tailored Qwen 3.6 in the IDE. More efforts like https://www.jetbrains.com/junie/ are needed for the majority persona to perceive value from changing their workflow again.
HN's "job" is better outcomes with less friction at each persona.
andyferris · 2026-08-26 12:53:40 UTC
In raw numbers of humans... "the early majority" surely would be those that use ChatGPT or Gemini (aka Google) and pay between $0 and $20 a month?
I would be surprised if the specialist that knows that various Chinese models exist and/or that a user might choose a harness and model separately are a "majority" even of the early variety... in terms of revenue, humans, tokens, or any metric.
(Happy to be proven wrong)
vintermann · 2026-08-26 12:12:18 UTC
> these models have no chance at end user penetration and loyalty until there's a single focused survivor.
This reminds me a lot of media horse-race reporting, saying that "candidate X has no chance unless they" and "candidate Y has a strong showing in", and it's very thinly cover for the publication liking Y and disliking X, avoiding talking about actual policy, and trying as much as they can to make their predictions self-fulfilling.
hypfer · 2026-08-26 12:12:56 UTC
> have no chance at end user penetration and loyalty until there's a single focused survivor.
But why does that matter? End users (I believe, feel free to correct) do not really contribute all that much revenue-wise.
They're certainly not the SOTA target audience.
The professional market doesn't need a household name. They need the most sensible tool for the job, and the CN models right now tick many boxes when it comes to that.
Comments
Seems legit.
It's really hard to know how good it is. So much hype around it.
https://x.com/davis7/status/2091285712566140986
Wenghi is behind DeepSWE, one of the best benchmarks.
1. Just variance in pass@K. If you prompt any model multiple times you'll see a large variance. N=1, but I find chinese open source models have a higher variance than higher-RL'd models like fable/opus.
2. They legitimately shipped a new RL checkpoint over the 7 days, which I find hard to believe.
I am leaning towards 1.
That's very plausible to have, identify, and fix in a day; especially when you get community feedback in the wild.
https://livebench.ai/
while here it outperforms Fable by a significant margin:
https://oxalpha.com/
but if the latter is true, will people still say it was "distilled" from Fable?
Many people and even software engineers fall for this all the time.
Most of these people are from crypto pivoting to AI doing this.
AI has made this easier and cheaper and it is going to get a LOT worse.
Imagine lots of websites with typosquatting and looking exactly the same as another website, vibe coded and cloned within seconds.
The public have no chance.
This one that is listed is designed to rank on Google as an informational source (although unofficial and not from z.ai which is why I said it is phishing)
Assuming you are technical you are able to discern this, imagine the average person.
No chance.
Yes, and its these sorts of websites I am asking about.
> This one that is listed is designed to rank on Google as an informational source (although unofficial and not from z.ai which is why I said it is phishing)
Again, what is there to phish?
The number one problem in LLM benchmarking is that people try to draw conclusions from sample sizes far too small to conclude anything but "it works sometimes, it fails sometimes, hard to say which is better." (The number two problem is that people run benchmarks blindly without checking that they measure something meaningful.)
Source: https://twitterwebviewer.com/?tweet=2091116504787935350
> 'GAD consistently surpasses standard sequence-level distillation, delivering superior generalization and achieving performance that rivals the proprietary teacher. These results validate GAD as an effective and robust solution for black-box LLM distillation.'
No RL, although I'm a little bit surprised to see MS Research publishing a paper on distilling GPT5?
> We reserve 500 samples of LMSYS-Chat-1M-Clean as the primary test set. We also include test datasets consisting of a 500-sample subset split from Dolly [6], the 252-sample SelfInst dataset [37], and the 80-question Vicuna benchmark [3] to evaluate out-of-distribution generalization. We report the GPT-4o evaluation scores [45, 10], where GPT-4o first generates reference answers and then scores the output of the student model against them. We also conduct human evaluations on the LMSYS-Chat-1M-Clean test set for qualitative assessment.
The metric used there is me screaming at my screen per operating hours.
Does it matter? IMO not really. Weights are open after all. (Or.. soon at least for 5.3)
And critically, like contracts in general, Anthropic's terms of service is only binding upon the user/counterparty. So even if a company say specifically sought out 'claude-like' content, and claude code traces available on the internet, if they don't use the Anthropic platform there is no ToS claim.
It is a capable small model, but it’s not frontier level. The interesting part will be seeing the model size, how it responds to quantization, and how fast it runs on the kind of non-server hardware that we can buy without selling a kidney.
They claim an "independent community benchmark" (pass-fail evaluation on tasks) here: https://oxalpha.com/ox-alpha-vs-fable-5
One thing that I think matters a lot for the non developers, all three of us (sol, oxa, and me) usually agree that oxa's write up is far far better. It explains the situation very well, and has great structure for its write ups. Sol gets the job done, but it's terrible at re-explaining the problem for humans, at laying out information. It also doesn't show it's thinking, so it's imo a terrible peer to work with!
on toy benches it made quite a few mistakes but was able to fix all of them on its own
(meaning more tokens, more turns, more tool calls — but same outcome as gpt 5.6 sol)
Didnt know they exist - looks very good, maybe even better than Archive.ph
Chinese labs are not releasing all of their model weights. Qwen is known as an open weight model by most, but their top model is not open weight.
Releasing weights is a marketing strategy for newer labs to get their brand out there.
https://mashable.com/article/chatgpt-gpt-4o-ai-retirement-pr...
> We should seize this rare, historic opportunity to encourage open source, openness, collaboration and sharing. [1]
People have pointed that this seemingly made Alibaba/Qwen turn around from closing their models (this was rumored after the shakeup early this year [2]) and release the weights for even the Max variant of their new models, which they previously did not.
1: http://english.scio.gov.cn/topnews/2026-07/18/content_118605...
2: https://simonwillison.net/2026/Mar/4/qwen/
In the end there's barely any moat that any of the ludicrously "valued" AI companies have - the only thing justifying the valuations of SpaceX/xAI/Tesla/Anthropic/OpenAI is a supposed "secret sauce" that, frankly, barely exists any more.
Everyone and their dog can go and run LLMs for dirt cheap on their own hardware.
Regular pricing is $0.15 input, $0.50 output... but currently 50% off, making it $0.075 input and $0.25 output. That beats most of the V4 Flash providers, but not all, and obviously tokens per task may not be equivalent.
I've also just noticed the blog post reveals the Artificial Analysis score - it's a 57, so it's Opus 4.8 / 5.6 Terra level.
https://z.ai/blog/glm-5.3-flash
Z.AI is the only provider for GLM 5.3 on OpenRouter. I don't see 5.3 on Hugging Face. Not sure if this new model is "full GLM" or something smaller, or if they will like Moonshot AI publish weights but put restrictive license [1], which will again leave Z.AI as single GLM model provider on OpenRouter.
[1] https://huggingface.co/moonshotai/Kimi-K3/blob/main/LICENSE
It took me a year talking about it until my wife knew that ChatGPT and Gemini are two different things.
PS: some replies, especially if you do a deep dive on comment history, clearly expose the joint effort to drum up support for Chinese models. This has been clear on HN lately as anything even slightly critical of Chinese tech gets downvoted unnaturally quickly. One can just wonder what's behind the effort...
Ox is just GLM. And z.ai is the maker of GLM.
The main players in the openweight model market have been known for a while.
And they already have significant user penetration.
Not to go off topic but I am pleased to see open model support from US companies like Poolside.ai, NVIDIA, IBM, Google, etc.
It took me a year talking about it until my wife knew that Kimi K3 and GLM 5.3 are two different things.
Fast follower persona clusters around emerging zeitgeist across the tellings. At the moment, arguably that's mostly Qwen for everyday hobbyists, and GLM for those that can run 512GB to 1.5TB of memory. This persona is seeking viable applied results: "I have frontier at home".
The early majority pick things up after models are curated into apps like LM Studio or one's platform app of choice, usually at least one major release behind because it takes that long to choose and package into mass distribution.
This is the step where early majority persona "has no idea" what the parade of weird names is about, they care about qualia of the conversations they try to have.
This persona is, at present, very under-served, and likely to remain so until mass devices can perform feeling like 27B at Q4 large quality better, or workplace devices can achieve a pragmatic utility like 135B at Q8 or better.
Harnesses that work where the workplace persona lives bridge this. This persona doesn't care the Chinese model name, they care "does it code?" For that, the applied harness and model take time to be matched, as JetBrains did harnessing a tailored Qwen 3.6 in the IDE. More efforts like https://www.jetbrains.com/junie/ are needed for the majority persona to perceive value from changing their workflow again.
HN's "job" is better outcomes with less friction at each persona.
I would be surprised if the specialist that knows that various Chinese models exist and/or that a user might choose a harness and model separately are a "majority" even of the early variety... in terms of revenue, humans, tokens, or any metric.
(Happy to be proven wrong)
This reminds me a lot of media horse-race reporting, saying that "candidate X has no chance unless they" and "candidate Y has a strong showing in", and it's very thinly cover for the publication liking Y and disliking X, avoiding talking about actual policy, and trying as much as they can to make their predictions self-fulfilling.
But why does that matter? End users (I believe, feel free to correct) do not really contribute all that much revenue-wise. They're certainly not the SOTA target audience.
The professional market doesn't need a household name. They need the most sensible tool for the job, and the CN models right now tick many boxes when it comes to that.