rotative
the decoder · September 22, 2026

Xiaomi's MiMo-V2.6-Pro leads open AI models, Anthropic claims Claude was used

xiaomi released its mimo-v2.6-pro model, leading open ai models at a low cost, while anthropic accuses xiaomi of misusing claude. this impacts the open ai model landscape.

read in 简体中文 · 繁體中文 · 日本語 · 한국어

xiaomi released its mimo-v2.6-pro model, which leads current open ai model rankings with a score of 46 on the intelligence index from artificial analysis. the model costs $0.435 per million input tokens and $0.87 per million output tokens, making it a fraction of the cost of similar models. xiaomi achieved this performance through expanded reinforcement learning, and released its tools and training tasks openly, including 7,000 ready-made training tasks with automatic graders. however, anthropic accuses xiaomi of improperly siphoning off data through the claude model to train its own model lineup, which xiaomi denies.

read at source ↗

← rotative