Xiaomi MiMo-V2.6-Pro / Flash
Top-ranked open-weights model on the Artificial Analysis Intelligence Index (46); Pro is 1.02T MoE, Flash is 310B, both MIT licensed.
Why it matters
World's top open-weights model with 1M-token omnimodal context, MIT license, and the full RL training framework published.
What you could build with it
Self-host MiMo-V2.6-Flash (MIT licensed) as a local fallback model for when hosted APIs are down or rate-limited.
Does it hold up?
Open and self-hostable, but hardware requirements are heavy; realistically more practical through hosted inference or serious infra than as a weekend build target.
Built with Xiaomi MiMo-V2.6-Pro / Flash
- MiMo-V2.6 community tests compilationyoutube · Compilation of 29 first-day community tests: island town, lava lamp, black hole, 3D Sonic, Blender villa, local Flash runs.
- Xiaomi's MiMo V2.6 Model Debuts With Unprecedented Efficiency Gains — Winbuzzerarticle · Independent review confirming Pro's 46-point lead atop the Artificial Analysis open-weights index and a $0.13 per-index-task cost, plus Flash-vs-Pro benchmark splits on DeepSWE and Terminal Bench 4.0.
- MiMo-V2.6-Pro takes the open weights crown — i-scoop.euarticle · Coverage of the $2.6M live-streamed RL run and MIT-licensed release, with API pricing ($0.435/$0.87 per M tokens) unchanged from V2.5.
Learn more
First spotted on article: source.
More AI models
ElevenLabs Eleven v4 + v4 Turbo: new emotive TTS architecture with a 100ms real-time variantElevenLabs released Eleven v4 and Eleven v4 Turbo on September 28: a new text-to-speech architecture with inline…models · JEV 0.72VoiceStudio: open-source, fully-local ElevenLabs alternative with voice cloning, dubbing and dictation in 646 languagesVoiceStudio is an open-source, fully-local voice AI studio: voice cloning from a 3-second sample, voice design, video…models · JEV 0.71DeepSeek V4-Flash official API: public beta with upgraded agent capabilities, Responses API and Codex supportDeepSeek's official V4-Flash is now a public beta API with massively upgraded agent capabilities - benchmark scores…models · JEV 0.69Mistral Forge: enterprise platform for training and continuously improving proprietary modelsMistral AI launched Forge (Sep 28) — a full-lifecycle model training platform (pre-training, SFT, DPO/ODPO, RL…models · JEV 0.69kev: open, trainable Jev-like family of small decision models on Qwen3.5/3.8A family of small decision models (0.8B to 27B) built on Qwen3.5/3.8 that reproduces Jev's prefill-only typed-decision…models · JEV 0.68GPT-Synopsys: OpenAI and Synopsys multi-year deal to build an AI model that operates EDA toolsOpenAI and Synopsys signed a multi-year agreement on Sep 30 to jointly develop GPT-Synopsys, a specialized model that…models · JEV 0.66
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.