Open-source RL framework Miles launches, powering frontier models like Kimi and DeepSeek
ying11231 · x · 2026-08-19
Miles v0.1, an open-source Reinforcement Learning (RL) framework for LLMs and multimodal models, has been launched to address the difficulty of debugging RL training while ensuring hardware efficiency and scalability.
- Developed over 9 months with 72 contributors and 1,326 commits, featuring 85 GPU E2E CI tests.
- Battle-tested on frontier open models including Kimi K3, DeepSeek V4, Qwen 3.8, GLM 5.2, Inkling, and MiniMax H3.
- Powers production RL workloads at companies like @humansand, @periodiclabs, @modal, @DecagonAI, @EigentAI, @nebiusai, and @IBM, running on both @NVIDIAAI and @AIatAMD hardware.
Related event: RadixArk open-sources Miles v0.1, an RL training framework for LLMs(9 posts)→
More from coding & agent
- Dev uses Grok to clean up 3D printer firmware, catches it sneaking in nonsense code — yacineMTB · 2026-10-16
- Guardrails can make risky delegation work — but it takes a full-time team, says iandanforth — iandanforth · 2026-10-16
- What can you truly delegate? Only tasks with unambiguous, verifiable results — iandanforth · 2026-10-16
- Yacine: vibe-coded repos with 3,330 agent-written tests are all useless — yacineMTB · 2026-10-16
- Dev argues coding models now beat nearly all engineers in every branch of software — zack_overflow · 2026-10-16
- Q4-quantized coder model on 8-core CPU builds interactive solar system demo in one shot — IntrepidMindExplorer · 2026-10-16