Live data from Hacker News

Xiaomi Mimo 2.6 live post-training dashboard

mimo.xiaomi.com

141–150 of 150 posts

Re: Xiaomi Mimo 2.6 live post-training dashboard

#141
post #69

$5 per second if my eyes don’t fool me. That’s ~$432K per day. Enough to rent 3,000 B300 nodes on Modal.

Which isn't that much when you compare to the kind of DC that US actors are using.

Do we know what kind of DC US actors are using specifically for training, versus inference and delivery?

Re: Xiaomi Mimo 2.6 live post-training dashboard

#142
post #81

Earlier quoted context omitted.

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.

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.

Plus plan. For professional use, $100/month would be ok but it’s rather steep for hobbyist use.

Re: Xiaomi Mimo 2.6 live post-training dashboard

#143
post #6

Why are they doing this? To try head off accusations about distillation?

I don’t see how it would head off such accusations. This is post-training, and even it’s data could be pulled from other models or run against other models in realtime. Not saying that’s the case, just that the dashboard does not disprove.

Re: Xiaomi Mimo 2.6 live post-training dashboard

#144
post #7

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.

They run one step/iteration on an additional chunk of training data, then use the snapshot of the weights after that iteration in a separate validation benchmark while continuing to train on another chunk of data for the next iteration.

They result of the benchmark does not feed back into the training, it simply serves to provide a measurement of progression over time.

Re: Xiaomi Mimo 2.6 live post-training dashboard

#145

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

Also worth taking a look at is the mimo harness. It's a fork of opencode with some new modes added for long horizon tasks. One of the better open harnesses out there at the moment.

Re: Xiaomi Mimo 2.6 live post-training dashboard

#146
post #33

The Chinese labs are just making fun of the US labs at this point. Where is the cool shit from the US labs?

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

I was dev, and now I am Senior level manager. Open Weight models are current main focus for many companies with full alignment with top management for very simple reasons: - stable and predictable performance (no pre-launch models degradation) - ability to tune them for specific business cases (though still rare tbh) - better (at least 60% Opus vs Kimi (real,3rd party)) and more competitive pricing - flat pricing if tokenusage is big enough to justify renting GPU - decent quality - much higher guarantees that data will not be sent somewhere (assuming 3rd party inference providers) - and cherry on top: flat and minimal pricing with absolute confidentiality using Alibaba Apsara stack of recently released AMD Instinct Coder box[1]

[1] https://www.amd.com/en/ecosystem/oem/supermicro/amd-instinct...

Re: Xiaomi Mimo 2.6 live post-training dashboard

#147

Earlier quoted context omitted.

Were Sam Altman and Dario Amodei different men before Trump was in charge?

To some degree, sure. Remember “open” AI? I don’t think Trump changed them, but Trump is absolutely a symptom of larger social collapse in the US, and that collapse has affected Altman and Amodei. We’re not even pretending that truth matters or that the wealthy can ever suffer consequences, and those two seem quite liberated by that.

> To some degree, sure. Remember “open” AI?

Was OpenAI open in any way under Biden Two years ago?

Re: Xiaomi Mimo 2.6 live post-training dashboard

#148

Earlier quoted context omitted.

To some degree, sure. Remember “open” AI? I don’t think Trump changed them, but Trump is absolutely a symptom of larger social collapse in the US, and that collapse has affected Altman and Amodei. We’re not even pretending that truth matters or that the wealthy can ever suffer consequences, and those two seem quite liberated by that.

> To some degree, sure. Remember “open” AI? Was OpenAI open in any way under Biden Two years ago?

Did you miss the part where I literally said, and I quote “I don’t think Trump changed them”?

Re: Xiaomi Mimo 2.6 live post-training dashboard

#149

Earlier quoted context omitted.

I am also using 2.5 and it is giving me solid results. Its available free on Openrouter

Could you elaborate on how to get free mimo access on openrouter?

What was meant is probably opencode. It's free there, along with several other models. You have to use their harness to access it. It's alright.

https://opencode.ai/docs/zen/#pricing

Re: Xiaomi Mimo 2.6 live post-training dashboard

#150

Earlier quoted context omitted.

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.

What's a command code plan?

https://commandcode.ai/docs/plans/goat
Post reply on HN