Show HN: 1-Bit Bonsai, the First Commercially Viable 1-Bit LLMs(prismml.com)
60 pointsby PrismML4 hours ago15 comments
  • jjcman hour ago
    1 bit with a FP16 scale factor every 128 bits. Fascinating that this works so well.

    I tried a few things with it. Got it driving Cursor, which in itself was impressive - it handled some tool usage. Via cursor I had it generate a few web page tests.

    On a monte carlo simulation of pi, it got the logic correct but failed to build an interface to start the test. Requesting changes mostly worked, but left over some symbols which caused things to fail. Required a bit of manual editing.

    Tried a Simon Wilson pelican as well - very abstract, not recognizable at all as a bird or a bicycle.

    Pictures of the results here: https://x.com/pwnies/status/2039122871604441213

    There doesn't seem to be a demo link on their webpage, so here's a llama.cpp running on my local desktop if people want to try it out. I'll keep this running for a couple hours past this post: https://unfarmable-overaffirmatively-euclid.ngrok-free.dev

    • najarvg20 minutes ago
      Thanks for sharing the link to your instance. Was blazing fast in responding. Tried throwing a few things at it with the following results: 1. Generating an R script to take a city and country name and finding it's lat/long and mapping it using ggmaps. Generated a pretty decent script (could be more optimal but impressive for the model size) with warnings about using geojson if possible 2. Generate a latex script to display the gaussian integral equation - generated a (I think) non-standard version using probability distribution functions instead of the general version but still give it points for that. Gave explanations of the formula, parameters as well as instructions on how to compile the script using BASH etc 3. Generate a latex script to display the euler identity equation - this one it nailed.

      Strongly agree that the knowledge density is impressive for the being a 1-bit model with such a small size and blazing fast response

      • jjcm16 minutes ago
        > Was blazing fast in responding.

        I should note this is running on an RTX 6000 pro, so it's probably at the max speed you'll get for "consumer" hardware.

      • najarvg14 minutes ago
        I must add that I also tried out the standard "should I walk or drive to the carwash 100 meters away for washing the car" and it made usual error or suggesting a walk given the distance and health reasons etc. But then this does not claim to be a reasoning model and I did not expect, in the remotest case, for this to be answered correctly. Ever previous generation larger reasoning models struggle with this
        • jjcm10 minutes ago
          [dead]
    • uf00lme22 minutes ago
      The speed is impressive, I wish it could be setup for similar to speculative decoding
    • adityashankar42 minutes ago
      here's the google colab link, https://colab.research.google.com/drive/1EzyAaQ2nwDv_1X0jaC5... since the ngrok like likely got ddosed by the number of individuals coming along
      • jjcm36 minutes ago
        Good call. Right now though traffic is low (1 req per min). With the speed of completion I should be able to handle ~100x that, but if the ngrok link doesn't work defo use the google colab link.
        • adityashankar34 minutes ago
          The link didn't work for me personally, but that may be a bandwidth issue with me fighting for a connection in the EU
    • hmokiguess20 minutes ago
      wow that was cooler than I expected, curious to embed this for some lightweight semantic workflows now
  • keyle9 minutes ago
    Extremely cool!

    Can't wait to give it a spin with ollama, if ollama could list it as a model that would be helpful.

  • volume_tech10 minutes ago
    the speed is not just about storage -- at 1-bit you are reading roughly 16x less data from DRAM per forward pass compared to FP16. on memory-bandwidth-constrained hardware that is usually the actual bottleneck, so the speedup scales pretty directly. the ac
  • _fwan hour ago
    What’s the trade-off? If it’s smaller, faster and more efficient - is it worse performance? A layman here, curious to know.
    • kvdveeran hour ago
      Their own (presumably cherry picked) benchmarks put their models near the 'middle of the market' models (llama3 3b, qwen3 1.7b), not competing with claude, chatgtp, or gemini. These are not models you'd want to directly interact with. but these models can be very useful for things like classification or simple summarization or translation tasks.

      These models quite impressive for their size: even an older raspberry pi would be able to handle these.

      There's still a lots of use for this kind of model

    • adityashankar24 minutes ago
      If you look at their whitepaper (https://github.com/PrismML-Eng/Bonsai-demo/blob/main/1-bit-b...) you'll notice that it does have some tradeoffs due to model intelligence being reduced (page 10)

      The average of MMLU Redux,MuSR,GSM8K,Human Eval+,IFEval,BFCLv3 for this model is 70.5 compared to 79.3 for Qwen3, that being said the model is also having a 16x smaller size and is 6x faster on a 4090....so it is a tradeoff that is pretty respectable

      I'd be interested in fine tuning code here personally

  • ariwilson7 minutes ago
    Very cool and works pretty well!
  • alyxyaan hour ago
    I expect the trend of large machine learning models to go towards bits rather than operating on floats. There's a lot of inefficiency in floats because typically they're something like normally distributed, which makes the storage and computation with weights inefficient when most values are clustered in a small range. The foundation of neural networks may be rooted in real valued functions, which are simulated with floats, but float operations are just bitwise operations underneath. The only issue is that GPUs operate on floats and standard ML theory works over real numbers.
  • syntaxing2 hours ago
    Super interesting, building their llama cpp fork on my Jetson Orin Nano to test this out.
  • Archit3chan hour ago
    Doesn't Jevons paradox dictate larger 1-bit models?
  • hatthew40 minutes ago
    I feel like it's a little disingenuous to compare against full-precision models. Anyone concerned about model size and memory usage is surely already using at least an 8 bit quantization.

    Their main contribution seems to be hyperparameter tuning, and they don't compare against other quantization techniques of any sort.

  • yodon2 hours ago
    Is Bonsai 1 Bit or 1.58 Bit?
    • woadwarrior012 hours ago
      1-bit g128 with a shared 16-bit scale for every group. So, effectively 1.125 bit.
  • OutOfHerean hour ago
    How do I run this on Android?
    • najarvg7 minutes ago
      Pocket Pal is what I've seen used before. Although recently heard about "Off Grid" but not read any reviews about it or tried it personally so caveat emptor. Will see if the community has other suggestions
  • stogot2 hours ago
    What is the value of a 1 bit? For those that do not kno
    • jacquesm2 hours ago
      That you can process many operations with a single instruction.
    • SwellJoe2 hours ago
      0 or 1
      • jjcman hour ago
        Technically not in this case, or not effectively. The 0 or 1 correspond to a FP16 scaling factor for each group of 128 bits. The value fluctuates between each group of 128.
    • trebligdivad2 hours ago
      Speed and density.
  • 3 hours ago
    undefined
  • techpulselab18 minutes ago
    [dead]
  • simian19832 hours ago
    [dead]