Xiaomi Mimo 2.6 live post-training dashboard

(mimo.xiaomi.com)

459 points | by krackers 15 hours ago

35 comments

  • joelwallis 15 hours ago
    I been using MiMo-V2.5 to do most of my work as software engineer, on a variety of projects I'm working on, and I been VERY happy with ROI. The model is very powerful! Not perfect – I've run in hallucination loops once or twice, but nothing a stop-then-continue wouldn't solve.

    The cost is unbelievably low, and the quality of intelligence I get is equivalent to when I was working mostly with Anthropic models (late last year/early this year). I'm fully invested in MiMo and I'm very happy with it.

    -- PS: I also check almost daily to see if other models are capable of doing such great work. And they do – DS4F is powerful and DS41 is impressive, GLM 5.3 Flash gets a job done well, etc. – but when I add cost of M-token in the ROI math, Jeez! MiMo is an order of magnitude better.

    • ehsankia 3 hours ago
      > late last year/early this year

      That's an eternity when it comes to coding models.

      In my personal experience, we've had almost a step change every ~3 months this year, at least for bigger one-shot tasks. For example looking at Gemini Flash 3.0 vs 3.5 vs 3.8, it went 5% -> 30% -> 75% on DeepSWE, all since the start of the year.

      • epolanski 3 minutes ago
        In one sense you're right.

        In another one, Opus 4.6 level already solved 90% of my work-day tasks, so while better models have been instrumental into handling a higher % that does not mean that defaulting on cheaper models can't be good.

        I run DS 4.1 flash daily, and then cross check with gpt-6-astra and I've nuked 90% of my AI monthly bill while having higher limits and better performance/intelligence than I did just at the beginning of this summer.

      • pelagicAustral 1 hour ago
        tbf, I the happiest I've been working with claude is late last year/early this year (before March)...
        • epolanski 2 minutes ago
          That's because Opus 4.6 was the last good assistant model.

          Everything after it might be more "intelligent" but is super tuned around end-to-end task (and related benchmarks), not to act as an assistant.

          Now it's *you* being the assistant, reviewer, etc.

    • miyuru 6 hours ago
      Same here. It’s the first AI provider I actually gave money to, since they offered the model for free with a Mimo code for the first month or so, and it was great.

      These days, there are more intelligent models like DS4.1, but Mimo is very obedient, so I plan things with another model and give the implementation to Mimo.

    • rapind 10 hours ago
      I’ve been very pleased with DS 4.1 flash. Not so much the 4.0 models, but for coding (Rust) it’s been great so far (3 solid days of work).

      I’ll give Mimo a try.

      • trollbridge 10 hours ago
        MiMo is my backup whenever DeepSeek is down, had the price bump, is slow, etc.

        UltraSpeed was absolutely awesome. I miss it.

        DS 4.1 Flash is amazing. Well worth the extra cost.

    • alwinaugustin 12 hours ago
      I am also using 2.5 and it is giving me solid results. Its available free on Openrouter
      • rahmatawaludin 1 hour ago
        Could you elaborate on how to get free mimo access on openrouter?
    • walrus01 14 hours ago
      I've found that mimo v2.5 works for very basic things like a python script to do one thing, but it also is very 'dumb' compared to qwen 3.8-flash-next (I think the benchmark scores for terminal and coding specific benches back this up). And definitely not in the same class as like a GLM5.2 or 5.3. It's fast but makes basic mistakes that only get caught later.
      • girvo 9 hours ago
        The fact I can run Qwen 3.8 Flash Next locally, forever (on my DGX Spark-alike) is genuinely shocking to me. It’s crazy good for how small it is. Fast, too.
        • jonsoft 7 hours ago
          I made this 3D game in a day on the same setup with Qwen Code as agent: https://games.jonathanpage.com/

          And I am not a web developer! It's an extraordinary model.

          (Mouse and keyboard required)

        • walrus01 9 hours ago
          Yeah, I'm guessing you have a variant that fits in <128GB with 262k context? I have the unsloth Q8 GGUF of it here in a setup that with full context and ton of extra llama-server "--cache-ram" sits around 200GB RAM usage on a 256GB system, it's probably the best thing I've found for a 256GB class machine. Enough headroom for a rope/yarn extension to 524288 context if I need it.
          • girvo 8 hours ago
            Yep, the engrams are on NVMe (the speed penalty was lower than I expected) and it is quantised to fit.

            It’s good enough that I’m considering a second spark, or selling this and buying an M5 Ultra with 256GB for it

    • baxtr 5 hours ago
      Could you elaborate on how you check daily? Do you swap models for certain tasks?
    • jwpapi 14 hours ago
      May I ask why you ended up there instead of just using the heavy subsidized subscription. I’m actually curious.
      • eli 12 hours ago
        Mimo has subsidized subscriptions too
    • flexagoon 12 hours ago
      How does it compare with DS 4.1 Flash in your experience, if you ignore the cost?
    • james2doyle 15 hours ago
      2.5 Pro or the regular 2.5?

      I always found that those Mimo models to be really good at tool calling and following instructions

    • ignoramous 5 hours ago
      > GLM 5.3 Flash gets a job done well, etc. – but when I add cost of M-token in the ROI math, Jeez! MiMo is an order of magnitude better.

      API may be expensive, but I do 900m tokens (95% cached, ~0.4% output) on Z.ai's $18/mo coding plan with GLM 5.3 Flash.

      • miroljub 5 hours ago
        I wouldn't call that inexpensive.

        For comparison, I am currently at 6.6B tokens, 95% of monthly quota on a 10$ command code plan, mostly using DeepSeek flash 4.1, or some of the free models for easier tasks.

    • esafak 14 hours ago
      How fast is it compared with the other Chinese models?
      • ricardobeat 13 hours ago
        They both are in the 50-100 tok/s range. The Mimo v2.5 Pro Ultraspeed beta could reach 1000 tok/s, hoping they can do something similar for the new model, it was amazing.
    • wangxili1997 3 hours ago
      [flagged]
    • electroglyph 12 hours ago
      [flagged]
      • NuclearPM 12 hours ago
        Real?
        • electroglyph 9 hours ago
          mimo 2.5 has been a big underperformer since shortly after it's release imo. i cancelled my sub after the first month. purposefully using 2.5 right now is just handicapping yourself for no reason.
          • NuclearPM 8 hours ago
            I understand now. You used the wrong word.
    • yeeeloit 14 hours ago
      [flagged]
      • senordevnyc 14 hours ago
        Yeah, this Brazilian dude who has been a contributor here on HN longer than your anonymous account is shilling for a Chinese model company. Makes sense.
      • platinumrad 14 hours ago
        Are you accusing them of astroturfing? Why is it strange for someone to say something topical?
  • dr_dshiv 13 hours ago
    Well, if open source AI is dangerous (for OpenAI/Anthropic IPOs?), this is like watching a time bomb.
    • dzonga 12 hours ago
      the open burial started when zAI served their latest model on all Chinese chips.

      now we r just noticing the grave getting dug deeper.

    • skybrian 11 hours ago
      For my own usage, Luna is cheap enough that I don't care if other models are cheaper. I'm interested if another model is in some way better and not too expensive.
      • rapind 10 hours ago
        Luna is great but makes a lot of mistakes at high and lower in my experience (large rust codebase). I use Luna Max for asynchronous subagent reviews and am very happy with its work, but it’s slow af.
      • ijidak 9 hours ago
        What plan are you on?

        Trying to understand why users are using Luna when Sol seems essentially unlimited on the pro plan. Unless you have jobs running 24/7.

        • teki_one 8 hours ago
          Sol is useless atm on the Plus plan, 1-2 questions 5-10m to get through the 5h allowance. (used to be good, can change any day)
    • SlightlyLeftPad 10 hours ago
      I think it might be closer to this:

      https://www.debtdefaultclock.us/

  • passive 13 hours ago
    Neat! I've been trying out their next model for the last week, which I assume is a version of this, and it's been a good experience so far.

    I had used 2.5-pro for a hefty chunk of development, and found it to work like a somewhat forgetful senior engineer who was new to my project. Very capable, would almost always choose a reasonable option, if not always the best one for the project, and not great at multi-tasking. Generally, made me comfortable not scrutinizing the code line-by-line, but still needed a bit of steering once projects got to a reasonable size.

    The next model is a clear step up in the multi-tasking capability at least, with me very rarely having to steer the implementation of a well-defined issue. In terms of code, I found MiMo-V.2.5-pro to be extremely conservative, implementing minimal solutions. The next model seems a little bit more ambitious, in positive ways, making good guesses about gaps/next steps. It also seems to be a fair bit better at design, at least for the little bit I've done, it was good at translating my concepts to practical elements on screen, and cleaned things up nicely as I made suggestions.

  • ricardobeat 13 hours ago
    For reference, Mimo-v2.5-Pro scored 19% on DeepSWE 1.1. This is looking great.

    Fable scores 70%, Kimi K3 69%, Astra 74% (all on max effort).

    https://deepswe.datacurve.ai/blog/deepswe-v1-1

    • Cookingboy 11 hours ago
      2.6-pro just reached 63.7% by step 10, it's on step 11 right now.

      Even flash reached 60.7% by step 12, and it's on step 16 now.

      This is so exciting lmao.

      • arcanemachiner 2 hours ago
        DeepSWE is saturated now IMO, and is basically worthless. Lots of new models get around 74%. Shame too, because it was a pretty decent benchmark for a few months there.
        • brookst 3 minutes ago
          It is saturated, but that doesn’t mean worthless. Seeing 72% is low-signal, but 30% is still meaningful.
    • markasoftware 3 hours ago
      gemini 3.8 flash is also 74% and google just started letting all their engineers use claude...go figure
      • ehsankia 2 hours ago
        > and google just started letting all their engineers use claude

        That's misleading.

        1. Having different models available is useful for A/B testing and helping improve Gemini itself.

        2. They have an enterprise offering for Antigravity (their agentic coding platform), and they need to test that it works well with non-Gemini models too.

  • krm01 15 hours ago
    This is pretty neat. What would be a good reason for the other Model providers to not do this?
    • kibae 15 hours ago
      Speculating here, but I assume researchers can make a reasonable estimate of the size of closed models based on factors like training time, training speed, and the number of tokens processed.

      Also, Anthropic and OpenAI probably want to keep each other on their toes so they don’t end up on the wrong side of another Opus 4.6 / GPT-5.3-Codex situation, where one lab releases a model only for the other to drop a better one hours later.

      • Bolwin 6 hours ago
        I don't really remember a situation, which of those models supposedly beat the other?

        I still opus 4.6 though not for code

      • jwpapi 14 hours ago
        I think first of all it’s not an obvious idea, also the marketing surplus for other providers is not as big for openai/anthropic as for xiaomi and last but not least I’m pretty sure you can withdraw methodology from here.

        I’m saying who has a million dollars for me, so I can make my own model?

    • nikcub 8 hours ago
      this is remarkable transparency in an otherwise hyper competitive and secretive industry
  • liuliu 15 hours ago
    When you run benchmarks while training, isn't that the definition of contamination? Asking because I am not sure if this is normal in big labs now.
    • jampekka 15 hours ago
      Kinda yes. The benchmarks become part of the validation set, which means the models get slightly overfit to them if they are used as criteria for stopping the training. But a lot less compared to using them in the training data.

      I'd guess everybody uses at least some benchmarks as stopping criteria, which is kinda sensible, but it also does induce some benchmaxxing, and explains partly why the newest models always tend to eke out in benchmarks.

      https://en.wikipedia.org/wiki/Training,_validation,_and_test...

      • liuliu 15 hours ago
        Correct. If just stopping criteria, that is less contaminated. The question gets muddier once you also use it to determine hyperparameters during small-scale runs.
    • lucrbvi 15 hours ago
      They are using it to evaluate checkpoints during the training, they are probably not using the benchmarks for training the models. It's a common practice for big reinforcement learning runs.
    • nodja 13 hours ago
      They exist to detect degradation. Datasets are not perfect and if a batch contains too much bad data it can ruin a run, also an opportunity to find bad data and improve the dataset filtering.
    • SwellJoe 15 hours ago
      You gotta have something to aim at. And, presumably, the benchmark is not part of the training data, it is the test against which the model is tested at each stage; is behavior moving in the right direction?
    • esafak 14 hours ago
      Not if you don't train against them.
      • kingstnap 13 hours ago
        It's implicitly trained against. There is like information leakage with researchers messing with the training parameters and checkpoints used.

        It's not the direct feedback loop of RL but its not far.

  • fzysingularity 13 hours ago
    Very cool to see the openness here, and likely more like this will come from smaller startups where they win users on transparency.
  • ProfessorLayton 15 hours ago
    2.6 Pro: >started 2026-09-15 10:32 UTC

    For some reason I thought training took much, much longer than what the progress bar suggests.

    This is really neat, I'm currently using mimo 2.5 pro, and it's decent (or great given the price). Hopefully their next one is multimodal.

    • GaggiX 15 hours ago
      These are post-training reinforcement learning steps.
      • krackers 15 hours ago
        Yes, updated the submission title to say "post-training" to hopefully prevent further confusion
  • thehamkercat 15 hours ago
    This is crazy, but sadly anthropic/openai will never do this, what has happened to this world, where chinese companies are more open than US or even EU companies
    • b3lvedere 3 hours ago
      Is that a bad thing?
    • medlazik 14 hours ago
      Neoliberalism, that famously open and transparent economic ideology
      • atemerev 6 hours ago
        Ah, one Donald Trump, a famous neoliberal.
        • stymaar 2 hours ago
          Were Sam Altman and Dario Amodei different men before Trump was in charge?
  • ttul 10 hours ago
    $5 per second if my eyes don’t fool me. That’s ~$432K per day. Enough to rent 3,000 B300 nodes on Modal.
    • stymaar 2 hours ago
      Which isn't that much when you compare to the kind of DC that US actors are using.
  • ssn2000 7 hours ago
    Total run cost is $1.2M until now, what resources are they using to train their model? Wish they shared more details on that and what the MFU metrics are.
  • speedgoose 15 hours ago
    I didn't know 2 thirds of the training data would be source code.
    • jerrygenser 15 hours ago
      that is the the "data used to improve the model" when signing up for the subscription plans
    • leothetechguy 15 hours ago
      this is the rl run, not the pretraining run
      • ahmadyan 14 hours ago
        even in pre-training, usually 30%-50% is code these days.
        • leothetechguy 2 hours ago
          That would be far too high in my opinion. But happy if anybody can give insights from their own experience with pretraining runs.
  • rao-v 10 hours ago
    I absolutely love that someone is doing this! Why isn’t IBM for Granite or Google for Gemini?

    If you are going to develop a near frontier model, and you don’t think you have special sauce up your sleeve, why not making training runs and RL environment scores etc. visible to the world?

    I’m genuinely learning quite a bit just from the dashboard

    • gtirloni 9 hours ago
      They think they have the special sauce. Even if they do, what would they get in return for doing that?
  • kkotak 8 hours ago
    Wouldn't us observing this break down the model superposition and make it dumber? :)
  • jstummbillig 6 hours ago
    Wow, spending money on training an almost-frontier-model is much more time intensive than I thought it was.
  • rozab 15 hours ago
    Why are they doing this? To try head off accusations about distillation?
    • bayindirh 15 hours ago
      Sometimes you're confident about what you're doing and show how you work to the world.

      Keeping the garage door open, or at least making the door translucent. It's always cool.

    • jampekka 14 hours ago
      That China's official policy is now to prefer open models and open model development may be a part of it.
      • culi 14 hours ago
        BRICS just had a New Delhi meeting where Xi pushed a 5-point plan on AI cooperation that centered on open source models
      • Aboutplants 14 hours ago
        With that policy in place, labs might be incentivized to be creative in their openness. This being fun/free PR
    • anemic 14 hours ago
      Bottom of the page says "Open is what we value."
  • wolttam 15 hours ago
    Hah, it would be great to see more labs pick this up.
  • ernsheong 13 hours ago
    Mino 2.5 has been my workhorse for coder and tester agents (the ones planner agents delegate tasks to)
  • monneyboi 3 hours ago
    Refreshing, now let's make this a default feature. I imagine a "Upcoming models" list with links to these kind of dashboards.
  • wg0 5 hours ago
    "Slow down this much openness in AI or we won't get our trillion dollars valuations!"

    Google had this GPT long go and a wise man within Google noted:

    "We don't have any maot neither does anyone else."

    The AI bubble burst is guaranteed and is only delayed by IPOs.

    • user43928 3 hours ago
      Nothing is guaranteed.

      Open models have not yet caught up with February's Mythos checkpoint.

      Meanwhile OpenAI is solving millennium problems, and their compute is still fully utilized.

      • wg0 2 hours ago
        "Stealing millennium problems" would be more complete if not accurate description. And that 99.99% of the market is not interested in solving millennium problems is the other fact.
    • b3lvedere 3 hours ago
      I wonder what we will do with the discarded data centers and its hardware..
      • Ylpertnodi 1 hour ago
        Copper can be stole, but that's already in progress.
  • thenews 9 hours ago
    been using the 2.5 mimo for side projects, works amazing
  • dr_kiszonka 10 hours ago
    Very curious that everyone here (so far) seems to assume this dashboard presents real data.
    • hsbalanxvxjsmab 9 hours ago
      Haha yeah pretty wild how easily you can see the data is fake by the repeating numbers (refresh the page the progress goes back in time constantly) + watch for restarts. They say they happen but 0 data correlates the log messages. Just a replay of old data or being fed by an llm so they convince people they are open
      • Bolwin 6 hours ago
        The intermediate tickers are fake but real data comes in and resets it. Its like a progress bar essentially. We don't call progress and bars fake
        • hsbalanxvxjsmab 1 hour ago
          I do when their fake like this site is. Insane people blindly believe this stuff
  • esafak 14 hours ago
    That's the kind of transparency we need! That DeepSWE benchmark puts it in frontier territory: https://artificialanalysis.ai/agents/coding-agents?coding-ag...
  • Alifatisk 5 hours ago
    Can we call this open AI?
  • singularity2001 1 hour ago
    "Claude Distill Requests":'hidden'
  • alescalaios 2 hours ago
    [dead]
  • tcbbd 8 hours ago
    [dead]
  • heronbank 10 hours ago
    [dead]
  • Toslink 13 hours ago
    [dead]
  • sinuhe69 5 hours ago
    [flagged]
  • pppkin 9 hours ago
    [dead]
  • impulser_ 14 hours ago
    The Chinese labs are just making fun of the US labs at this point.

    Where is the cool shit from the US labs?

    • culi 14 hours ago
      With other software, devs convince their managers of the importance of using open source stuff in their stack. With AI, it's usually managers choosing what models to use for the devs. The US labs don't need to give a damn how much devs like open source
      • impulser_ 13 hours ago
        This isn't about liking open source. This is about the labs just being cool and doing cool shit instead of the opposite which is Anthropic where all they talking about is killing everyone and taking everyone's job.
        • dlisboa 10 hours ago
          These labs are still (for the time being) made of people, who reflect their lives onto the work.

          The US population is much more pessimistic and doomsday driven these days, whereas the Chinese are more optimistic and future driven.

      • noir_lord 13 hours ago
        > The US labs don't need to give a damn how much devs like open source

        In the short term, true.

        In the long term, unknown but typically when you hold progress that way while other countries don't you at best end up becoming siloed while the rest of the world continues on without you.

    • hsbalanxvxjsmab 9 hours ago
      You mean all of the frontier models that the Chinese distillation clones are copying? Yeah kinda cool imo. If a dashboard showing training for a model that doesn't even come close to anything us labs have released in 6 months is "cool", then you're a loser
      • bicepjai 8 hours ago
        Hahaha. Is that Sam or Dario with throwaway account. This sounds like calling social security, a free handout. Who distills the distillaters? Get it?
      • impulser_ 6 hours ago
        Why the fuck would you or I care about that?

        Anthropic and OpenAI literally stole from every human in history and youre out here complaining that the Chinese are distilling models and releasing them to the public?

        Why do you care?

        • hsbalanxvxjsmab 1 hour ago
          Because without those labs to distill from the pathetic Chinese labs wouldn't have anything. Im not impressed by them copying US labs not sure why you are. But go off ccp bot
      • atemerev 6 hours ago
        No crying in the copyright casino.
  • dude250711 12 hours ago
    Distillation in real-time? Very interesting!
    • Cookingboy 11 hours ago
      That "training cost" is just live revenue count for Anthropic/OpenAI API calls!

      /s

  • hsbalanxvxjsmab 9 hours ago
    This is so very clearly fake? See the message stating the flash 2.6 flash run was restarted and 0 graphs correlate that restart
    • Retro_Dev 7 hours ago
      A restart of the process does not necessarily mean reverting the model state. I don't know why you would even do that, because you'd lose all the progress you made.
      • hsbalanxvxjsmab 1 hour ago
        It said restarted step 15 5 mins ago and the progress showed they were working on step 16 for a day
  • levocardia 15 hours ago
    You'd think they would make it less obvious that they are running their whole operation with Claude
    • ricardobeat 13 hours ago
      If you're thinking of the UI style, definitely not Claude. It is incapable of writing a clear sentence like "what each step's samples are made of", would have used all-caps for everything, more padding and gradients.
      • conception 9 hours ago
        I hope this is /s because it’s very easy to get Claude to write sensibly. That’s why AI slop writing is so annoying because it’s so easy to avoid with any amount of effort at all.
        • ricardobeat 2 hours ago
          In my experience Opus and Sonnet 5 subtly ignore most instructions related to writing style, and continue to sound the same half of the time. Do you have a successful skill/prompt to share?
    • jambutters 11 hours ago
      They'd be running in the red then cause they charge way less than Claude. Sorry but it just doesn't make logical sense. They have open source, papers, and self hosting too
    • SwellJoe 15 hours ago
      It's not obvious to me. What's the tell?
    • iammrpayments 5 hours ago
      Did you come here to astroturf or are you a big fan of Claude
    • cpcabbge 5 hours ago
      Do tell cause I can't