Which harness are u folks using, I have tried opencode and claude code. Both absolutely keep hanging due to the model running into loops and becoming unavailable.
Unable to do even simple things
I had Ox Alpha working on coding tasks for a couple days non-stop, via OpenRouter and OpenCode Zen. Crush harness. It was able to complete tasks at a level that I'd put between Sonnet and Opus. It makes few mistakes, but is not that smart.
The main issue for me, is that it degraded into a doom loop several times. One of them was running the same bash command about a thousand times. The last model I've used that had this problem was Mimo 2.5, which is quite dated at this point. As a result of this, you cannot leave it unattended / not usable for agents.
I couldn't get past all the network errors on OpenCode. Seemed smart enough, and was useful when I was low on usage on Claude, but beyond that, really hard for me to say whether it was Good or Bad.
Doom loops are as much of a model problem as it is deficiency of the harness. I have not seen any other open source harness that deals with them except the one I started because of this obvious gap.
See my other comment with examples where 0x Alpha is working non-stop on various projects with zero problems.
I usually see doom loops when working with quants. Likely theyre trying to maximize the viability of a efficient model quant that can bw upgraded. Like cutting coke to get crack, quantiry over quality.
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.
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 comment sections).
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.
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.)
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.
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
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.
"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?
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.
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.
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.
It writes pretty clean code and holds context alright, but it starts stumbling and losing the plot on complex bash scripts with pipelines. Waiting for the weights to drop so we can dig under the hood and see what is going on there
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.
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.
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.
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.
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.
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.
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
> 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.
I’m more curious on the size. If it’s smaller than or equal size to GLM 5.3, this would be a crazy good model. If it’s closer to deepseek pro, it would be a good model. If it’s near Kimi K3, I think it’s competitive but nothing particularly differentiating.
Definitely agree. If it is small (eg. Qwen 3.8 28b or gpt-oss-120) then this might be amazing. If it is anywhere near Kimi K3 it would need to have some other differentiating factor than intelligence.
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.
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
> 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.
Apparently someone working at a 3rd party inference provider also got confused and posted confirmation about it being a glm-flash model, despite having an embargo on that info. Someone jumped in the comments and told them they missed the timezone :)
In any case it should be releasing in a few hours. Timezones are hard.
Rather than a pelican, for fun I showed it a couple of screenshots from Niu Lai and asked it to create an SVG inspired by the images. I explained a little about how the movie had been made by a mother & son team, initially derided but then went on to surprise cult box office success. It came up with this:
Created in a single turn - but technically not a "one-shot", because I gave it a tool to convert SVG to PNG so it could visualize what it had made. I asked it to keep iterating with tools during the same turn until it was happy.
I've also been using Ox Alpha for tasks that better resemble real work, and I'm really enjoying working with it. I've downgraded my Anthropic account so I can put some budget towards Ox Alpha instead, with the rumors that this one is going to be cheap. Opus & Fable are still better at getting large tasks / features done autonomously, but Ox Alpha can work autonomously too, and it's fun. I'm enjoying working with Ox in a way that I'm just not enjoying talking to the 5.0 Anthropic models. (As much as I don't want to say that, as someone with Claude /stickers on their laptop.)
Yeah, that's a fair point. "More intelligible than Adriano Celentano in Prisencolinensinainciusol" is not a high bar.
As another comparison, I went back to MiniMax M3 for a while last night. It was significantly faster than Ox, but I felt M3's replies were harder to parse, not quite getting to the point. But I guess I could curb that with some prompts.
It depends if the Ox Alpha pricing is as cheap as was being rumored. If it's competitive with DeepSeek Flash and significantly undercutting Luna, that feels like it will be significant.
Ox alpha is better at UI than GPT 5.6 Sol. Not a high bar considering Sol sucks at UI, but as someone who just has a codex sub, I've used almost 1B tokens of ox alpha these last few days to complement Sol smartness.
Inference was atrocious in terms of speed and constant timeouts. If it's served fast it will be a delight to use.
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...
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.
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.
> 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.
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.
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.
> 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.
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.
Only reason people are interested is it’s free at the moment. I wasn’t impressed by its performance. Once the model gets a price tag it’s usage will be negligible.
Calling it now: The big deal about this model is the sheer volume they were offering through openrouter and OpenCode. How? Chinese AI accelerators / nvidia-free stack
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.
The main issue for me, is that it degraded into a doom loop several times. One of them was running the same bash command about a thousand times. The last model I've used that had this problem was Mimo 2.5, which is quite dated at this point. As a result of this, you cannot leave it unattended / not usable for agents.
See my other comment with examples where 0x Alpha is working non-stop on various projects with zero problems.
An amusing thought of returning to your workstation to find it as an obsidian block after it gets stuck executing "dd" thousand times.
It sounds like you were using a quant model.
Very impressive model.
Here are some examples, open-source documented and the data available in HF datasets:
https://openzot.github.io/whetstone/ - https://github.com/openzot/whetstone
https://openzot.github.io/arcade/ - https://github.com/openzot/arcade
https://openzot.github.io/machinery/ - https://github.com/openzot/machinery
Didnt know they exist - looks very good, maybe even better than Archive.ph
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?
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.
Source: https://twitterwebviewer.com/?tweet=2091116504787935350
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.
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.)
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.
> '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.
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
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.
Related PR: https://github.com/jeffhajewski/latticedb/pull/5
The session used ~100K input tokens, ~60K output tokens, and ~80K thinking tokens.
I reviewed it using gpt-sol-medium, and it seems to be satisfied with it's work.
Seems legit.
It's really hard to know how good it is. So much hype around it.
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://x.com/davis7/status/2091285712566140986
Wenghi is behind DeepSWE, one of the best benchmarks.
Where? And "Tonight" in which timezone?
Apparently someone working at a 3rd party inference provider also got confused and posted confirmation about it being a glm-flash model, despite having an embargo on that info. Someone jumped in the comments and told them they missed the timezone :)
In any case it should be releasing in a few hours. Timezones are hard.
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)
https://x.com/syneryder/status/2091978367579156569/photo/1
Created in a single turn - but technically not a "one-shot", because I gave it a tool to convert SVG to PNG so it could visualize what it had made. I asked it to keep iterating with tools during the same turn until it was happy.
I've also been using Ox Alpha for tasks that better resemble real work, and I'm really enjoying working with it. I've downgraded my Anthropic account so I can put some budget towards Ox Alpha instead, with the rumors that this one is going to be cheap. Opus & Fable are still better at getting large tasks / features done autonomously, but Ox Alpha can work autonomously too, and it's fun. I'm enjoying working with Ox in a way that I'm just not enjoying talking to the 5.0 Anthropic models. (As much as I don't want to say that, as someone with Claude /stickers on their laptop.)
That’s very valid, but right now every other model I use is easier to talk to than Opus 5.0
Opus 5.0 has an impenetrable way of communicating. I can parse it, but it takes so much more work than it should.
As another comparison, I went back to MiniMax M3 for a while last night. It was significantly faster than Ox, but I felt M3's replies were harder to parse, not quite getting to the point. But I guess I could curb that with some prompts.
It depends if the Ox Alpha pricing is as cheap as was being rumored. If it's competitive with DeepSeek Flash and significantly undercutting Luna, that feels like it will be significant.
hard agree. it does not really feel "smart", but the personality is super refreshing
Inference was atrocious in terms of speed and constant timeouts. If it's served fast it will be a delight to use.
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...
It took me a year talking about it until my wife knew that Kimi K3 and GLM 5.3 are two different things.
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.
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.
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.
Not to go off topic but I am pleased to see open model support from US companies like Poolside.ai, NVIDIA, IBM, Google, etc.
[0]: https://openrating.io/blog/current-state-of-ai-model-fingerp...