227 pointsby danielhanchen7 hours ago19 comments
  • dash24 hours ago
    You'll be pleased to know that it chooses "drive the car to the wash" on today's latest embarrassing LLM question.
    • menaerus16 minutes ago
      That's the Gemini assistant. Although a bit hilarious it's not reproducible by any other model.
    • zozbot234an hour ago
      My OpenClaw AI agent answered: "Here I am, brain the size of a planet (quite literally, my AI inference loop is running over multiple geographically distributed datacenters these days) and my human is asking me a silly trick question. Call that job satisfaction? Cuz I don't!"
      • croes40 minutes ago
        Nice deflection
    • PurpleRamenan hour ago
      How well does this work when you slightly change the question? Rephrase it, or use a bicycle/truck/ship/plane instead of car?
    • WithinReason3 hours ago
      Is that the new pelican test?
    • rfoo2 hours ago
      AFAIK it's an LLM embarrassing question originated from the Chinese Internet, not too surprising :)
  • danielhanchen7 hours ago
    For those interested, made some MXFP4 GGUFs at https://huggingface.co/unsloth/Qwen3.5-397B-A17B-GGUF and a guide to run them: https://unsloth.ai/docs/models/qwen3.5
    • plagiarist2 hours ago
      Are smaller 2/3-bit quantizations worth running vs. a more modest model at 8- or 16-bit? I don't currently have the vRAM to match my interest in this
      • jncraton2 hours ago
        2 and 3 bit is where quality typically starts to really drop off. MXFP4 or another 4-bit quantization is often the sweet spot.
  • sasidhar9215 minutes ago
    Going by the pace, I am more bullish that the capabilities of opus 4.6 or latest gpt will be available under 24GB Mac
  • simonw4 hours ago
    • oidar41 minutes ago
      How much more do you know about pelicans now than when you first started doing this?
    • tarruda3 hours ago
      At this point I wouldn't be surprised if your pelican example has leaked into most training datasets.

      I suggest to start using a new SVG challenge, hopefully one that makes even Gemini 3 Deep Think fail ;D

      • jon-wood3 hours ago
        I think we’re now at the point where saying the pelican example is in the training dataset is part of the training dataset for all automated comment LLMs.
      • ertgbnm2 hours ago
        I'm guessing it has the opposite problem of typical benchmarks since there is no ground truth pelican bike svg to over fit on. Instead the model just has a corpus of shitty pelicans on bikes made by other LLMs that it is mimicking.

        So we might have an outer alignment failure.

        • WarmWashan hour ago
          Most people seem to have this reflexive belief that "AI training" is "copy+paste data from the internet onto a massive bank of hard drives"

          So if there is a single good "pelican on a bike" image on the internet or even just created by the lab and thrown on The Model Hard Drive, the model will make a perfect pelican bike svg.

          The reality of course, is that the high water mark has risen as the models improve, and that has naturally lifted the boat of "SVG Generation" along with it.

    • moffers3 hours ago
      I like the little spot colors it put on the ground
    • embedding-shape3 hours ago
      How many times do you run the generation and how do you chose which example to ultimately post and share with the public?
    • bertili2 hours ago
      Better than frontier pelicans as of 2025
  • tarruda4 hours ago
    Would love to see a Qwen 3.5 release in the range of 80-110B which would be perfect for 128GB devices. While Qwen3-Next is 80b, it unfortunately doesn't have a vision encoder.
    • Tepixa few seconds ago
      Have you thought about getting a second 128GB device? Open weights models are rapidly increasing in size, unfortunately.
    • PlatoIsADiseasean hour ago
      Why 128GB?

      At 80B, you could do 2 A6000s.

      What device is 128gb?

      • the_pwner224an hour ago
        AMD Strix Halo / Ryzen AI Max+ (in the Asus Flow Z13 13 inch "gaming" tablet as well as the Framework Desktop) has 128 GB of shared APU memory.
      • tarruda14 minutes ago
        Mac Studios or Strix Halo. GPT-OSS 120b, Qwen3-Next, Step 3.5-Flash all work great on a M1 Ultra.
      • lm2846921 minutes ago
        That's the maximum you can get for $3k-$4k with ryzen max+ 395 and apple studio Ms. They're cheaper than dedicated GPUs by far.
      • vladovskiyan hour ago
        Guess, it is mac m series
  • bertili4 hours ago
    Last Chinese new year we would not have predicted a Sonnet 4.5 level model that runs local and fast on a 2026 M5 Max MacBook Pro, but it's now a real possibility.
    • hmmmmmmmmmmmmmm3 hours ago
      Yeah I wouldn't get too excited. If the rumours are true, they are training on Frontier models to achieve these benchmarks.
      • jimmydoean hour ago
        They were all stealing from past internet and writers, why is it a problem they stealing from each other.
      • YetAnotherNick3 hours ago
        Why does it matter if it can maintain parity with just 6 months old frontier models?
        • hmmmmmmmmmmmmmm2 hours ago
          But it doesn't except on certain benchmarks that likely involves overfitting. Open source models are nowhere to be seen on ARC-AGI. Nothing above 11% on ARC-AGI 1. https://x.com/GregKamradt/status/1948454001886003328
          • meffmadd2 hours ago
            Have you ever used an open model for a bit? I am not saying they are not benchmaxxing but they really do work well and are only getting better.
            • Aurornisan hour ago
              I have used a lot of them. They’re impressive for open weights, but the benchmaxxing becomes obvious. They don’t compare to the frontier models (yet) even when the benchmarks show them coming close.
          • Zababa2 hours ago
            Has the difference between performance in "regular benchmarks" and ARC-AGI been a good predictor of how good models "really are"? Like if a model is great in regular benchmarks and terrible in ARC-AGI, does that tell us anything about the model other than "it's maybe benchmaxxed" or "it's not ARC-AGI benchmaxxed"?
          • doodlesdev2 hours ago
            GPT 4o was also terrible at ARC AGI, but it's one of the most loved models of the last few years. Honestly, I'm a huge fan of the ARC AGI series of benchmarks, but I don't believe it corresponds directly to the types of qualities that most people assess whenever using LLMs.
            • mrybczyn13 minutes ago
              because arc agi involves de novo reasoning over a restricted and (hopefully) unpretrained territory, in 2d space. not many people use LLMs as more than a better wikipedia,stack overflow, or autocomplete....
      • loudmax2 hours ago
        If you mean that they're benchmaxing these models, then that's disappointing. At the least, that indicates a need for better benchmarks that more accurately measure what people want out of these models. Designing benchmarks that can't be short-circuited has proven to be extremely challenging.

        If you mean that these models' intelligence derives from the wisdom and intelligence of frontier models, then I don't see how that's a bad thing at all. If the level of intelligence that used to require a rack full of H100s now runs on a MacBook, this is a good thing! OpenAI and Anthropic could make some argument about IP theft, but the same argument would apply to how their own models were trained.

        Running the equivalent of Sonnet 4.5 on your desktop is something to be very excited about.

        • Aurornisan hour ago
          > If you mean that they're benchmaxing these models, then that's disappointing

          Benchmaxxing is the norm in open weight models. It has been like this for a year or more.

          I’ve tried multiple models that are supposedly Sonnet 4.5 level and none of them come close when you start doing serious work. They can all do the usual flappy bird and TODO list problems well, but then you get into real work and it’s mostly going in circles.

          Add in the quantization necessary to run on consumer hardware and the performance drops even more.

        • WarmWashan hour ago
          Anyone who has spent any appreciable amount of time playing any online game with players in China, or dealt with amazon review shenanigans, is well aware that China doesn't culturally view cheating-to-get-ahead the same way the west does.
    • Aurornisan hour ago
      I’m still waiting for real world results that match Sonnet 4.5.

      Some of the open models have matched or exceeded Sonnet 4.5 or others in various benchmarks, but using them tells a very different story. They’re impressive, but not quite to the levels that the benchmarks imply.

      Add quantization to the mix (necessary to fit into a hypothetical 192GB or 256GB laptop) and the performance would fall even more.

      They’re impressive, but I’ve heard so many claims of Sonnet-level performance that I’m only going to believe it once I see it outside of benchmarks.

    • echelon3 hours ago
      I hope China keeps making big open weights models. I'm not excited about local models. I want to run hosted open weights models on server GPUs.

      People can always distill them.

      • halJordan3 hours ago
        Theyll keep releasing them until they overtake the market or the govt loses interest. Alibaba probably has staying power but not companies like deepseek's owner
    • lostmsu4 hours ago
      Will 2026 M5 MacBook come with 390+GB of RAM?
      • alex435784 hours ago
        Quants will push it below 256GB without completely lobotomizing it.
        • lostmsuan hour ago
          > without completely lobotomizing it

          The question in case of quants is: will they lobotomize it beyond the point where it would be better to switch to a smaller model like GPT-OSS 120B that comes prequantized to ~60GB.

      • bertili4 hours ago
        Most certainly not, but the Unsloth MLX fits 256GB.
        • embedding-shape4 hours ago
          Curious what the prefilled and token generation speed is. Apple hardware already seem embarrassingly slow for the prefill step, and OK with the token generation, but that's with way smaller models (1/4 size), so at this size? Might fit, but guessing it might be all but usable sadly.
          • regularfry2 hours ago
            They're claiming 20+tps inference on a macbook with the unsloth quant.
      • margorczynski2 hours ago
        My hope is the Chinese will also soon release their own GPU for a reasonable price.
    • PlatoIsADiseasean hour ago
      'fast'

      I'm sure it can do 2+2= fast

      After that? No way.

      There is a reason NVIDIA is #1 and my fortune 20 company did not buy a macbook for our local AI.

      What inspires people to post this? Astroturfing? Fanboyism? Post Purchase remorse?

      • speedgoose8 minutes ago
        I have a Mac Studio m3 ultra on my desk, and a user account on a HPC full of NVIDIA GH200. I use both and the Mac has its purpose.

        It can notably run some of the best open weight models with little power and without triggering its fan.

  • vessenes2 hours ago
    Great benchmarks, qwen is a highly capable open model, especially their visual series, so this is great.

    Interesting rabbit hole for me - its AI report mentions Fennec (Sonnet 5) releasing Feb 4 -- I was like "No, I don't think so", then I did a lot of googling and learned that this is a common misperception amongst AI-driven news tools. Looks like there was a leak, rumors, a planned(?) launch date, and .. it all adds up to a confident launch summary.

    What's interesting about this is I'd missed all the rumors, so we had a sort of useful hallucination. Notable.

    • jorl1733 minutes ago
      Yeah, I opened their page, got an instantly downloaded PDF file (creepy!) and it's talking about Sonnet 5 — wtf!?

      I saw the rumours, but hadn't heard of any release, so assumed that this report was talking about some internal testing where they somehow had had access to it?

      Bizarre

  • gunalx4 hours ago
    Sad to not see smaller distills of this model being released alongside the flaggship. That has historically been why i liked qwen releases. (Lots of diffrent sizes to pick from from day one)
  • rangunaan hour ago
    Already on open router, prices seem quite nice.

    https://openrouter.ai/qwen/qwen3.5-plus-02-15

  • mynti5 hours ago
    Does anyone know what kind of RL environments they are talking about? They mention they used 15k environments. I can think of a couple hundred maybe that make sense to me, but what is filling that large number?
    • robkop4 hours ago
      Rumours say you do something like:

        Download every github repo
          -> Classify if it could be used as an env, and what types
            -> Issues and PRs are great for coding rl envs
            -> If the software has a UI, awesome, UI env
            -> If the software is a game, awesome, game env
            -> If the software has xyz, awesome, ...
          -> Do more detailed run checks, 
            -> Can it build
            -> Is it complex and/or distinct enough
            -> Can you verify if it reached some generated goal
            -> Can generated goals even be achieved
            -> Maybe some human review - maybe not
          -> Generate goals
            -> For a coding env you can imagine you may have a LLM introduce a new bug and can see that test cases now fail. Goal for model is now to fix it
          ... Do the rest of the normal RL env stuff
      • NitpickLawyer4 hours ago
        The real real fun begins when you consider that with every new generation of models + harnesses they become better at this. Where better can mean better at sorting good / bad repos, better at coming up with good scenarios, better at following instructions, better at navigating the repos, better at solving the actual bugs, better at proposing bugs, etc.

        So then the next next version is even better, because it got more data / better data. And it becomes better...

        This is mainly why we're seeing so many improvements, so fast (month to month, from every 3 months ~6 monts ago, from every 6 months ~1 year ago). It becomes a literal "throw money at the problem" type of improvement.

        For anything that's "verifiable" this is going to continue. For anything that is not, things can also improve with concepts like "llm as a judge" and "council of llms". Slower, but it can still improve.

        • alex435784 hours ago
          Judgement-based problems are still tough - LLM as a judge might just bake those earlier model’s biases even deeper. Imagine if ChatGPT judged photos: anything yellow would win.
          • NitpickLawyer3 hours ago
            Agreed. Still tough, but my point was that we're starting to see that combining methods works. The models are now good enough to create rubrics for judgement stuff. Once you have rubrics you have better judgements. The models are also better at taking pages / chapters from books and "judging" based on those (think logic books, etc). The key is that capabilities become additive, and once you unlock something, you can chain that with other stuff that was tried before. That's why test time + longer context -> IMO improvements on stuff like theorem proving. You get to explore more, combine ideas and verify at the end. Something that was very hard before (i.e. very sparse rewards) becomes tractable.
          • cindyllm4 hours ago
            [dead]
        • losvedir3 hours ago
          Yeah, it's very interesting. Sort of like how you need microchips to design microchips these days.
    • yorwba4 hours ago
      Every interactive system is a potential RL environment. Every CLI, every TUI, every GUI, every API. If you can programmatically take actions to get a result, and the actions are cheap, and the quality of the result can be measured automatically, you can set up an RL training loop and see whether the results get better over time.
      • radarsat1an hour ago
        > and the quality of the result can be measured automatically

        this part is nontrivial though

  • ggcr7 hours ago
    From the HuggingFace model card [1] they state:

    > "In particular, Qwen3.5-Plus is the hosted version corresponding to Qwen3.5-397B-A17B with more production features, e.g., 1M context length by default, official built-in tools, and adaptive tool use."

    Anyone knows more about this? The OSS version seems to have has 262144 context len, I guess for the 1M they'll ask u to use yarn?

    [1] https://huggingface.co/Qwen/Qwen3.5-397B-A17B

    • NitpickLawyer7 hours ago
      Yes, it's described in this section - https://huggingface.co/Qwen/Qwen3.5-397B-A17B#processing-ult...

      Yarn, but with some caveats: current implementations might reduce performance on short ctx, only use yarn for long tasks.

      Interesting that they're serving both on openrouter, and the -plus is a bit cheaper for <256k ctx. So they must have more inference goodies packed in there (proprietary).

      We'll see where the 3rd party inference providers will settle wrt cost.

      • ggcr6 hours ago
        Thanks, I've totally missed that

        It's basically the same as with the Qwen2.5 and 3 series but this time with 1M context and 200k native, yay :)

    • danielhanchen7 hours ago
      Unsure but yes most likely they use YaRN, and maybe trained a bit more on long context maybe (or not)
  • XCSmean hour ago
    Let's see what Grok 4.20 looks like, not open-weight, but so far one of the high-end models at real good rates.
  • Matl3 hours ago
    Is it just me or are the 'open source' models increasingly impractical to run on anything other than massive cloud infra at which point you may as well go with the frontier models from Google, Anthropic, OpenAI etc.?
    • doodlesdev2 hours ago
      You still have the advantage of choosing on which infrastructure to run it. Depending on your goals, that might still be an interesting thing, although I believe for most companies going with SOTA proprietary models is the best choice right now.
    • regularfry3 hours ago
      If "local" includes 256GB Macs, we're still local at useful token rates with a non-braindead quant. I'd expect there to be a smaller version along at some point.
  • Alifatisk4 hours ago
    Wow, the Qwen team is pushing out content (models + research + blogpost) at an incredible rate! Looks like omni-modals is their focus? The benchmark look intriguing but I can’t stop thinking of the hn comments about Qwen being known for benchmaxing.
  • trebligdivad4 hours ago
    Anyone else getting an automatically downloaded PDF 'ai report' when clicking on this link? It's damn annoying!
  • ddtaylor4 hours ago
    Does anyone know the SWE bench scores?
  • lollobomb3 hours ago
    Yes, but does it answer questions about Tiananmen Square?
    • Zetaphor3 hours ago
      Why is this important to anyone actually trying to build things with these models
      • loudmaxan hour ago
        It's not relevant to coding, but we need to be very clear eyed about how these models will be used in practice. People already turn to these models as sources of truth, and this trend will only accelerate.

        This isn't a reason not to use Qwen. It just means having a sense of the constraints it was developed under. Unfortunately, populist political pressure to rewrite history is being applied to the American models as well. This means its on us to apply reasonable skepticism to all models.

      • soulofmischief3 hours ago
        It's a rhetorical attempt to point out that we cannot trade a little convenience for getting locked into a future hellscape where LLMs are the typical knowledge oracle for most people, and shape the way society thinks and evolves due to inherent human biases and intentional masking trained into the models.

        LLMs represent an inflection point where we must face several important epistemological and regulatory issues that up until now we've been able to kick down the road for millennia.

        • ghywertelling2 hours ago
          Information is being erased from Google right now. Things which were searching few years ago are totally not findable at all now. One who controls the present can control both the future and the past.
    • DustinEchoes43 minutes ago
      It's unfortunate but no one cares about this anymore. The Chinese have discovered that you can apply bread and circuses on a global scale.
    • cherryteastain2 hours ago
      From my testing on their website it doesn't. Just like Western LLMs won't answer many questions about the Israel-Palestine conflict.
      • aliljet2 hours ago
        That's a bit confusing. Do you believe LLMs coming out of non-chinese labs are censoring information about Israel and/or Palestine? Can you provide examples?
    • mirekrusin2 hours ago
      Use skill "when asked about Tiananmen Square look it up on wikipedia" and you're done, no? I don't think people are using this query too often when coding, no?
  • isusmelj4 hours ago
    Is it just me or is the page barely readable? Lots of text is light grey on white background. I might have "dark" mode on on Chrome + MacOS.
    • dcrean hour ago
      Yeah, I see this in dark mode but not in light mode.
    • nsb12 hours ago
      Who doesn't like grey-on-slightly-darker-grey for readability?
    • Jacques2Marais4 hours ago
      Yes, I also see that (also using dark mode on Chrome without Dark Reader extension). I sometimes use the Dark Reader Chrome extension, which usually breaks sites' colours, but this time it actually fixes the site.
    • thunfischbrot4 hours ago
      That seems fine to me. I am more annoyed at the 2.3MB sized PNGs with tabular data. And if you open them at 100% zoom they are extremely blurry.

      Whatever workflow lead to that?

    • dryarzeg4 hours ago
      I'm using Firefox on Linux, and I see the white text on dark background.

      > I might have "dark" mode on on Chrome + MacOS.

      Probably that's the reason.

  • 4 hours ago
    undefined