Add Xiaomi MiMo-V2.6-Pro-RL and MiMo-V2.6-Flash-RL

Xiaomi has released and open-sourced the new MiMo-V2.6 series. Please consider adding both MiMo-V2.6-Pro-RL and MiMo-V2.6-Flash-RL to the LiveBench leaderboard.

Model information

• Organization: Xiaomi MiMo
• Models:
  • MiMo-V2.6-Pro-RL
  • MiMo-V2.6-Flash-RL
• Architecture: Sparse Mixture-of-Experts (MoE)
• Context length: 1M tokens
• Modalities: Text, Image, Video, Audio
• Type: Open-weight multimodal reasoning / agentic models

Official links

• MiMo-V2.6 collection: https://huggingface.co/collections/XiaomiMiMo/mimo-v26
• MiMo-V2.6-Pro-RL: https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Pro-RL
• MiMo-V2.6-Flash-RL: https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL
• Xiaomi MiMo GitHub: https://github.com/XiaomiMiMo/MiMo

Both models are part of Xiaomi's latest MiMo-V2.6 release and show major improvements over the previous MiMo-V2.5 generation across coding and agentic benchmarks.

It would be great to have independent LiveBench evaluations for both models and compare them with other current frontier models.

Thanks!