“MiMo-V2.6 Scaling Reinforcement Learning Towards Self-Improvement”
This is a dev post classified by Jev as AI dev tools (a launch), kept by the Dev Radar because it carries real work, not commentary.
“MiMo-V2.6 Scaling Reinforcement Learning Towards Self-Improvement” Xiaomi MiMo just completed their largest RL scaling run so far, openly. And it came along with a technical report, outlining how they've scaled RL across compute, environments, and grading. tl;dr The run costs $2.6M for Pro and $0.9M for Flash, with ~44% spent on rollouts, 41-44% on training, and 14% on grading. Each RL step uses 1,568 prompts x 16 rollouts, producing ~25K trajectories and 2.7-3.7B tokens, with contexts up to 1M tokens. New techniques include groupwise agentic grading, which compares successful trajecto
Posted by alphaXiv (56.1k followers) 3 h ago · 49 likes · 2.5k views · view the original post on X. Kept by the Dev Radar as AI dev tools. Tools mentioned: alphaXiv.
More dev work like this
- The Dot platform has processed 6.18B tokens this week, representing 34.33% of our… — @usedotai
- What is Jev, the tool buzzing across Silicon Valley lately? — @hayden090807
- PSA - Claude code: Turn off the prompt suggestions, save ~10% of your limits/spend — @akshdeeps_001
- 智谱 GLM Coding Plan 开启中秋加国庆双节畅享活动。 — @0xLogicrw
- 😱 什么？！仅 0.6B 本地开源决策模型在 Typed Decisions 上压过 Laya ！ — @NFT_Chen
- Runs the 2.78-trillion-parameter Kimi K3 model on a single CPU using 8 GB of RAM without… — @tom_doerr
- CLAUDE.md 里写了一堆规则，Claude Code 照样有一半当没看见，这事很多人都碰到过。 — @GitHub_Daily
- 🤯卧槽！开源圈又杀出个比 Laya 更狠的 System-1 决策核模型 AgentJev-0.6B ！ — @NFT_Chen
Every post is read and classified by Jev (TypeSafe): what it is, which market it belongs to, and whether the link is a real tool. 20.7k posts from 4.9k X accounts over the last 21 days, 2.4k tools, 12 markets. Collected every 5 minutes, fully re-ranked every hour — last update 2026-09-22 08:21 UTC. Full method.