Skip to content

Dataset · XiaomiMiMo

RL Environments for LLM Agents

XiaomiMiMo/MiMo-V2.6-RL-oss

Environment suite for training LLM agents with checks in code, cybersecurity, web, and music.

What's inside

These are RL environments for training LLM agents across multiple domains: software engineering, vulnerability reproduction, knowledge work, visual web development, and symbolic music. Verification types include executable tests, rule checks, rubric-based judging, and visual grading. The card provides Docker images and training code; modalities include documents, images, and text.

What you can do with it

  • RL training for LLM agents
  • Evaluating agent skills in coding and cybersecurity
  • Reproducing environments for web and music tasks
Size
1K–10K rows
Downloads this month
260.7K
Likes
1.1K

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("XiaomiMiMo/MiMo-V2.6-RL-oss")