Pelicans for Pro: https://tools.simonwillison.net/markdown-svg-renderer?url=ht...
Qwen3.8-27b pelican was amazing on Mac.
Pro [2]:, 1.02T total / 42B activated parameters
And later they further cut Sol and Terra pricing by 20% (maybe only in the API) and Luna by 80%.
In fact Luna still outperformed DeepSeek Flash 4.1 in cost per task on Artificial Analysis when I last checked.
However, Luna is slightly less intelligent. I have a feeling that it's pretty dumb and prone to hallucination unless running at xhigh or max effort, where it somehow manages to work quite well.
I did not personally test the open weight models beyond the old Qwen 3.6 27B, which produced unusably bad results for me.
The competition is great, and I hope Chinese models will continue to force leading US labs to offer models at a low price point.
That said, I don't think the Chinese labs have anything over OpenAI and Anthropic when it comes to capability or efficiency - I have no reason not to believe the US labs have even lower cost to serve the models.
Do also remember China is this far in the AI race despite all chip restrictions from America. If they were in equal standards I truly think Chinese models would have long surpassed American ones. Also would like to remind how Anthropic CEO is being hostile and blaming Chinese models with distilling meanwhile their own models claimed to be Qwen¹ and their stance against open models is negative² and they still keep blaming China for it.
1- https://news.ycombinator.com/item?id=48671252
2-https://www.anthropic.com/news/position-open-weights-models
Given the difference in compute, it seems plausible.
However, the researchers at the US labs are surely no less talented, and they have better access to hire talent globally.
They too have to serve their models efficiently at a large scale, and with current capacity constraints this must be a top priority.
Why wouldn't he? If there really was 25,000 accounts breaking ToS any CEO would at minimum be upset. Evidence of Claude distilling qwen would be damning but that a) makes no sense b) doesn't exist afaik.
OpenAI reduced prices and Anthropic increased weekly usage limits.
One simple task: I needed an LLM to go through and clean up a few thousand page descriptions and titles in my personal search engine index, where the human web page authors had put in no effort sigh. I did a shoot out between Claude, Luna, GLM 5.3 Flash and Deepseek. Despite the high cost, Claude's descriptions were terrible, and even Opus warned me that the descriptions coming back from Haiku were "generalized, not accurate". I expected I would choose Luna because of price, and occasionally it did have wonderful descriptions (one captured emotion in a way no other model did). But in the end, the GLM 5.3 Flash descriptions were the easiest to read, they flow well while also being accurate & including necessary keywords, and being highly affordable. So it won out. It's a task that is nowhere near frontier, but a task where somehow China is better than frontier.
Maybe Terminal Bench 4.0 and ExploitGym are reasonable.
Terminal Bench 4.0
GPT 6 Astra 59.6
Claude Fable 5.1 55.1
Claude Opus 5 49.0
MiMo-V2.6-Pro 34.9
MiMo-V2.6-Flash 28.8
DeepSeek V4.1 Flash 26.8
MiMo-V2.5-Pro 1.5
GPT 6 Astra 42.4
Claude Fable 5.1 30.4
Claude Opus 5 22.1
MiMo-V2.6-Pro 17.8
MiMo-V2.6-Flash 6.0
MiMo-V2.5-Pro 0.1
DeepSeek V4.1 Flash 74.2
Claude Opus 5 74.0
GPT 6 Astra 74.0
MiMo-V2.6-Pro 71.9
Claude Fable 5 70.0
MiMo-V2.6-Flash 67.9
MiMo-V2.5-Pro 19.0Maybe I am being routed to more quantised versions or less capable models with system prompt to fake Astra or Fable.
Some features of the release I like:
- Demonstration of diverse tasks, such as using a DAW
- Graphs from various benchmarks and price ranges
- Real world use of the model in scientific environments
[1] https://sebastianraschka.com/llm-architecture-gallery/per-la...
[2]: See DS 4.1-Flash and Qwen-3.8-Next.
It’s an NVFP4 quant, but it fits, and is surprisingly capable.
(or is it somewhere else)
This one!
I'd recommend pointing your agent at it (after installing sparkrun), and asking it to research the absolute latest in TP=1 Flash-Next - mine grabbed particular vLLM nightlies and mods to improve performance, and it was well worth it.
Averages ~25-35tok/s which isn't bad for a first attempt.
It's because offpeak electricity is cheaper?
Funnily it's perfect if you are in the Pacific Time Zone because you can use it daytime 9am to 5pm
Just tried 2.6 flash on a really niche topic I specialise in and it has done a really good job. They’ve definitely polluted their training data with claudeslop, but looking past the slop there is a decent model.
And as these models get better the pace of training is quickly speeding up too.
This doesn't bode particularly well for anthropic/OAI after they go public.
so weird to acknowledge someone being on the front edge, but not name it
MiMo-V2.6-Flash-310B-A15B roughly GPT-5.6 Luna / Claude 4.9 according to benchmarks MiMo-V2.6-Pro-1.02T-A42B roughly GPT-5.6 Sol / Opus 5 according to benchmarks
Perhaps with IQ2 flash will run on 128G M5?
I'm sure they're doing all kinds of terrible things, like all major companies. I just can't help but like them. Also, this model looks great, and I'll give their subscription a shot next month.
I'm in Europe, and here the options for home appliances are usually German (e.g. Philips), Balkan (e.g. Gorenje), or Xiaomi. Xiaomi is the best by far, and it's honestly not even close.
Their home appliances are so rock-solid that they actually still surprise me. I've gone from, e.g., having to replace electric water kettles every six months to buying one from Xiaomi and never replacing it. (Nigh on three years now.)
I really have a very positive impression of them.
I ask because my wife has the 15T and the camera is better than my iPhone 17 Pro. And while toying around with it I didn't notice any bloat.
Plus hers support native split screen which I kinda need to multitask on the go.
I'm so pissed at how bad Siri is compared to her android phone that I'm thinking about selling the iPhone to get a Huawei Pura Ultra.
but now I got my "proof".
It looks like the "Frontier Line" to me, which is also often misinterpreted. frontier does not mean the best models. It means all models that are not strictly dominated, meaning in most cases: Not same price or cheaper and more intelligent.
I personally would like the word frontier to be used with more criterias: Open Weights, per use-case, etc etc. This would make model selection easier, but I understand it’s not an easy thing to do.
- Pareto efficiency/Pareto curves: Basically the convex hull of points along the edge of a graph, indicating the best tradeoff between the axes. This is what the post is talking about.
- Pareto principle: this is the 80/20 rule you're talking about
What you call "frontier line" is also called "Pareto frontier" https://en.wikipedia.org/wiki/Pareto_front
Your description of it is basically correct though
The realtime dashboard they shared during training (https://mimo.xiaomi.com/rl/) was an incredible learning and teaching tool for me, and they’ve been unusually comprehensive in sharing details about their methodology (check out that tech report - it's got lots of clever behind the scene tricks like Google or Deepseek writeups) and benchmark scores (even the stuff they didn’t do well on).
If you’re releasing an open model going forward, please consider offering the community more of this transparency!
whey they all singing the same tune. it make me question what is their real motives.
they are afraid of Chinese good enough LLM model killing their margin. we already have story about US companies switch some task to use cheaper Chinese model hosted on Neoclouds. this might be the same play that China have done with EV and solar panel. they flood the market with cheaper alternative and slowly gain market share.
Here is one really neat bit:
A cutting edge training idea is on-policy RL, basically, it's not enough to say "here is an end to end agentic sequence (including tool calls etc.) that is perfect" you want to say "here is a sequence you might actually have generated that turns out to be correct".
Basically, it's more training efficient for models to improve with small tweaks to what they already do than from some perfect oracular answer (if you've ever tried to each people new skills, you’ve probably noticed this too :) )
When you do that, you care about how far the model you are updating (improving) has deviated from the one being used to generate rollouts (agentic rollouts for hard problems can take hours with lots of tool calls, so you can't keep redeploying every slight improvement).
Lo and behold, the dashboard literally has:
partial/avg_staleness (likely the measure of how many micro iterations the "generate answers" model is behind the "improving based on the occasional right answer" model)
train_infer_diff/new_infer/kl (a more direct KL divergence based way of measuring how much the two models have diverged)
How cool is that?!
Really the only thing missing was dataset descriptions, the dashboard only had random IDs like "dataset-zrso". I guess it's their lawyers fault.