The open weight crown moved to a phone maker
the open weight crown moved to a phone maker: mimo v2.6 pro leads the october leaderboard at 75.5, ahead of qwen3.8 max at 72.1, with mit licensed weights and a price under half a dollar per million tokens. six months ago the same ranking was a two company race between american labs.
the market for intelligence is being reorganized by whoever ships weights cheap and early.
Context
heise online's report of September 22, 2026 says Xiaomi's MiMo-V2.6-Pro takes the lead among open-weight models in the Artificial Analysis Intelligence Index, ahead of GLM-5.3, Kimi K3 and DeepSeek V4.1. It says Xiaomi charges 0.435 dollars per million input tokens without cache hits and 0.87 dollars per million output tokens through its API, that the model is a mixture of experts design with 1.02 trillion total and 42 billion active parameters, and that Xiaomi provides the weights of all three new models under the MIT license on Hugging Face. BenchLM lists MiMo-V2.6-Pro at 75.49 out of 100 overall, ranked 11 of 783 models, as an open weight reasoning model with a 1M token context window.
The 75.5 matches BenchLM's 75.49 for MiMo-V2.6-Pro, and the MIT license and the under one dollar per million tokens prices match heise's report. The note's 72.1 for Qwen3.8 Max was not located. ModelCap's own index, a different scoring, lists Qwen3.8 Max (0902) at 82.1 against 81.3 for MiMo-V2.6-Pro, at 2.00 dollars input and 6.00 dollars output per million tokens, so which model leads depends on the leaderboard used, and the Qwen model's weight availability was not checked. Heise notes an independent Artificial Analysis evaluation of the smaller Flash model was not yet available, and Xiaomi's benchmark gains are its own. The claim that six months ago the ranking was a two company race between American labs was not examined. The market for intelligence is being reorganized is the author's line.
Watch next
- The leaderboard source for the 72.1 figure and an independent look at weights availability for Qwen3.8 Max.
Sources
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 4 October 2026 at 22:47 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →