Skip to content
AI IntelligenceMay 3, 2026AI Intelligence
Article

Xiaomi's new open-weight model MiMo-V2.5-Pro nearly matches Anthropic's Claude Opus 4.6 on coding benchmarks while using much less energy.

It can autonomously code for hours.

Data Cube AI EditorialSource: The Decoder
01

Source Brief

Xiaomi's new open-weight model MiMo-V2.5-Pro nearly matches Anthropic's Claude Opus 4.6 on coding benchmarks while using much less energy. It can autonomously code for hours.