MiMo-V2.6 streams massive RL run live: 2B tokens per step, details to be open-sourced
burny_tech · x · 2026-09-17
Xiaomi's MiMo team spent six months studying how far RL can scale, and MiMo-V2.6 is mid-RL-run right now — with the training livestreamed publicly.
What they scaled:
- Compute: 2B tokens per step, 1568 prompts × 16 rollouts, fully async
- Environments/harnesses: multi-task agentic RL mixed across multiple harnesses in one run
- Grader compute: agentic in-group credit assignment with test-case and rubric-based rewards
The team says it will open-source the details piece by piece over the coming weeks. Commenters note that just watching the metric visualizations is highly educational — an unusually transparent move from a Chinese frontier lab.
Related event: Xiaomi live-streams MiMo-V2.6 RL training, burning over $1.08M in 36 hours(14 posts)→
More from Models
- Anthropic's server-side rule reignites the case for owning open models over renting AI — alex_verem · 2026-11-02
- Epoch AI: open-weight GLM 5.3 exploited the patch gap to hack 6 Firefox releases at ~$1,000 each — rohanpaul_ai · 2026-11-02
- SemiAnalysis: safety classifiers sandbag scores of Opus 5.5 and GPT-6 Astra — morqon · 2026-11-02
- Liquid AI's 1.7B LFM3 becomes first local model to nail live fixtures lookup on a MacBook — NielsRogge · 2026-11-02
- Claude upgrade-page traffic rebounded to 17.6M in September, near May's 18.6M peak — firstadopter · 2026-11-02
- Open-source GLiNER-PII scrubs 13 identifiers from a discharge note in 550ms on GPU — MaziyarPanahi · 2026-11-02