MAI-Code-1-Flash

(microsoft.ai)

143 points | by EvanZhouDev 1 hour ago

25 comments

  • camelmel 16 minutes ago
    Huh, according to that model card this is a 137B total parameter model.

    Performance doesn't seem that good:

    - MAI-Code-1-Flash (137B-A5B) = 51% on SWE-bench pro

    - Qwen3.6-35B-A3B = 49.5% on SWE-bench pro (https://huggingface.co/Qwen/Qwen3.6-35B-A3B)

    They benchmark against Claude Haiku but Haiku is not good, it's worse than tiny open models you can run locally or via API at 10% the cost.

    • giancarlostoro 7 minutes ago
      The take away is that this model is a smaller model that competes with Haiku, I would hope they come out with a "Sonnet" competing model, then Opus. I have been wondering why Microsoft is kind of "sleeping" on offering models they themselves have made on Copilot, maybe it was part of their deal with OpenAI? Not sure.
      • minraws 3 minutes ago
        They did release, MAI-Thinking-1 to compete with Sonnet. Totally not sure why that isn't at the top here.
        • giancarlostoro 1 minute ago
          Good question, and I missed that entirely!
    • wetpaws 2 minutes ago
      [dead]
  • hmokiguess 23 minutes ago
    Does anyone actually uses these smaller models for coding? If so, how? I usually Opus everything. Is the play to plan/design/architect with a heavier model than delegate structured tasks to these smaller ones? Would appreciate to hear someone's opinion on having done and tested both paths.
    • linuxhansl 5 minutes ago
      I am using Opus 4.x at work, and these "smaller" (20-80bn, 3-4bn active) models at home. Unfortunately there is no comparison, yet (IMHO anyway).

      With Opus I can work, trust its designs, architecture suggestions, and code changes, even in a complex code base.

      The smaller models seem to "try". They work for smaller tasks, but for more complex task it's often more work than doing it myself.

      I wish it were different, and maybe in a year or two it will be.

    • killermouse0 16 minutes ago
      I was wondering the same. I guess it makes sense to use a heavy weight model to make the entire design and split the work so that smaller models (possibly local one?) would then do the coding... But how would I even do that? I'm using Claude Code. Would I need support for this within the harness ?
    • ojr 19 minutes ago
      I use Gemini 3 Flash, I've seen the Claude Code setups, bullish on Anthropic people are driving up tokens but I am able to produce outcomes with a fraction of the money.
      • hmokiguess 16 minutes ago
        Do you mind sharing your workflow? What do you mean by fraction of the money, in my case personally, I'm yet to reach a session limit on the subscription plan. I'm not "tokenmaxxing" as they say, so hard to see a scenario in which the plan is expensive for the value I get.
        • dist-epoch 13 minutes ago
          If you don't hit a limit running Opus, it means you are very much in the loop.

          For example you probably don't have days where you ask Opus to review your whole code base and look for code duplication/technical debt/robustness issues, and then to fix some of the found issues, and do this 3-5 times until no big issues are found anymore.

          • hmokiguess 2 minutes ago
            What’s your prompt for this, the way you described it made it seem like there’s a generalizable way I can go about this. I just rely on a testing pipeline instead so can’t think of why I would need to proactively find holes where tests haven’t already done that for me.
    • newusertoday 17 minutes ago
      plan using opus execute using local
  • capten 44 minutes ago
    It's so weird to me that the benchmarks remain so low, but the models are marketed as revolutionary. And if you say that low coding capabilities aren't a problem, say that to the token price hike and 'general use' model setup.

    Why not sell it as a math agent? Why do I have to set up 4 agents to check each others' work?

    • redrove 26 minutes ago
      It’s about bang for buck. That high a score for 5B params is pretty good, nigh unbelievable a short while ago.

      It is my belief that smaller models will get better and better, and even cloud SOTA models will shrink.

      Yet another reason the current buildout will feel like the railroads.

      • Flere-Imsaho 19 minutes ago
        Yeah the future is probably a number of highly specialised small models you can run on your own hardware rather than massive frontier models in the cloud.

        That's what I'm betting on anyway.

        • search_facility 6 minutes ago
          MOE basically work that way already, QWEN/etc with low active params (A-number in name) allows to inference big models locally (only active params have to fit into memory)
        • thewebguyd 13 minutes ago
          That seems to be what Microsoft is betting on also based on what was shown at the BUILD keynote today + that new surface ultra and the surface mini PC with the new Nvidia chip. Nadella really played up local AI as the main use case they have in mind.
      • dist-epoch 6 minutes ago
        The SOTA models will not shrink, because the problems will get bigger, from "write me a C compiler" to "clone Stripe business and run it".
  • AntiRush 59 minutes ago
    The introductory blog post has a lot more information

    https://microsoft.ai/news/introducingmai-code-1-flash/

    and the model card

    https://microsoft.ai/pdf/MAI-Code-1-Flash-Model-Card.PDF

    The broader announcement of 7 MAI models seems to be where the 5B active in the title comes from

    https://microsoft.ai/news/building-a-hillclimbing-machine-la...

    • dang 15 minutes ago
      Thanks! I've changed the top link to the blog post and put the other links in the toptext.
  • mentos 30 minutes ago
    Shouldn’t the next model focus not be on code but system design?

    Seems like the work from a good system design to code is practically solved.

    Now it’s a matter of the design of the system. Or is that represented in these evals?

    • dist-epoch 9 minutes ago
      Have you tried system design with LLMs? I find them pretty good at suggesting 5 architectures for a problem and then iterating on the solutions.

      Even if I had no idea, going with the default suggestion would not be a terrible mistake, assuming you did describe your requirements relatively well.

  • efields 13 minutes ago
    Please test your websites in Safari. Almost all of your iOS users use it by default, and the desktop experience is pretty close to the mobile experience, so testing is easy.

    That scroll effect is jank city for me (yeah yeah works fine in Chrome/Edge).

  • tosh 39 minutes ago
    not open weight or at least I did not find anything indicating open weight
  • OsrsNeedsf2P 1 hour ago
    So it's trained on the SWE Bench Pro evalset
  • onlyrealcuzzo 51 minutes ago
    Gemma 4 26B-A4B scored exceptionally well with 20% less params, so this isn't unprecedented.
  • mmaunder 7 minutes ago
    You lost me at forced scrolling. Ugh!
  • gslepak 23 minutes ago
    Would be cool if this were an open model.
  • striking 21 minutes ago
    To be clear about the size of the model: MAI-Code-1-Flash is 137B A5B.
  • bguberfain 35 minutes ago
    It is good to se big companies like Microsoft launching LLMs. They have large amount of compute power and good scientists to create useful models.
    • ComputerGuru 26 minutes ago
      Microsoft has been releasing LLMs for years.
      • lemonish97 11 minutes ago
        They were mostly distilled or fine-tuned OAI models.
      • jwitthuhn 22 minutes ago
        And occasionally un-releasing them like with WizardLM.
      • ipsum2 24 minutes ago
        Sort of. Phi models were just trained on GPT outputs though.
  • kylehotchkiss 7 minutes ago
    "superintellegence team"

    Why not assign them to make windows good :D

  • hootz 44 minutes ago
    I'd love to see a tokens per second metric. I always prioritize speed over raw intelligence for flash models.
    • throwaw12 21 minutes ago
      > I always prioritize speed over raw intelligence for flash models.

      This model might have a perfect speed:

          for i in range(100):
            print(random.choices(words))
  • jMyles 12 minutes ago
    I'd really like to get back to an autocomplete flow, ideally with some shared and optimized context with the relationship with my larger agent models.

    But it seems like, by and large, even the faster models are now aimed at longer-running agentic flows and not sub-1s autocomplete. Or am I wrong about that?

  • ajyoon 43 minutes ago
    Scroll wheel hijacked on this entire domain
    • grav 5 minutes ago
      Fix:

        (() => {
        const KILL = ['wheel', 'mousewheel', 'DOMMouseScroll', 'touchmove'];
        const block = e => e.stopImmediatePropagation();
        for (const t of KILL) {
          window.addEventListener(t, block, { capture: true, passive: true });
          document.addEventListener(t, block, { capture: true, passive: true });
        }
        document.documentElement.classList.remove('lenis','lenis-smooth','lenis-scrolling','lenis-stopped');
        console.log('Scroll hijack disabled — native scrolling restored.');
        })();
    • matchbok3 38 minutes ago
      Yeah this website is horrendous to use. What were they thinking?
    • infraredshift 36 minutes ago
      [dead]
  • freediddy 36 minutes ago
    is 51% good enough to reliably use? There's no world in which I use an AI agent where it gets even 15% of the code wrong, that's as bad a Tesla FSD where you need to pay attention to the road while engaging FSD. What's the point? My attention is what I'm trying to relieve, not mostly correct functionality. The only thing that matters is whether you can one-shot code like Claude or Codex, I'm not interested in a small but mostly-okay-but-annoyingly-buggy-every-now-and-then AI.
    • VygmraMGVl 31 minutes ago
      Claude opus 4.6 scores 51.9% on the same benchmark. Microsoft's result is quite good.
  • LoganDark 24 minutes ago
    "Clean data" is impossible. Language models have polluted the landscape to such a degree it's impossible to filter them out now. OpenAI has no doubt discarded or muddled their dataset that was used to train the original ChatGPT, so there may be no dataset in existence now that isn't contaminated.
  • zb3 10 minutes ago
    So it's not an open model while not being much better? Meh.
  • ghord 9 minutes ago
    [dead]
  • vancekai 25 minutes ago
    [flagged]
  • Ozzie-D 26 minutes ago
    [dead]
  • fooker 59 minutes ago
    [flagged]
  • mattlondon 34 minutes ago
    Comparing against Claude 4.5? Aren't we up to 4.8? But disingenuous?
    • klardotsh 30 minutes ago
      They're comparing to Haiku, not Opus. Haiku is currently at 4.5.

      Even if it were Opus, comparing to a version number makes for an interesting snapshot of time comparison: if you knew how a model performed at whatever time in was in vogue, you can say "well, it looks like Model X is about 6 months/1 year/etc. behind the frontier SOTA" - which is exactly the discussion that happens in the open-weight/local LLM space. (interesting, MAI-Code-1-Flash does not appear to be such an open-weight model, following the western trend of locking models up)

    • 0vermorrow 33 minutes ago
      Latest Haiku (smallest Anthropic Model) is version 4.5, they haven't released a new version, hence the comparison to that.