Revision History

Auto-saved before update

Correction: dated a stale "at the time of writing" ranking claim and added MiMo-V2.6-Pro's Artificial Analysis Intelligence Index v5.0 score (30, 29 October 2026)

Verification fixes: board denominator

Update: Agent Arena #1 open model and Pareto frontier (Arena, 15-16 Oct 2026) with net-improvement explainer; MOPD checkpoints and tool-call repetition fix (27 Sep); Vals AI environment audit; Alignment Index and Code Arena results; Batch API pricing and V2.5 deprecation; Artificial Analysis currency check

Improve internal links (2026-10-13)

Update: expand the music demo from Xiaomi's launch post and 22 Sep X post; refresh Artificial Analysis rank/speed to 25 Sep; add trade-press reception (VentureBeat, SiliconANGLE); add HySparse2 successor-architecture section (Luo Fuli 23 Sep, arXiv 2609.26368)

Link CodeMidas; add Artificial Analysis cost per Intelligence Index task ($0.13) and its Pareto-frontier statement

Correction: comparison-model scores in Xiaomi's table are not all Xiaomi's own measurements (several match OpenAI's GPT-6 Astra launch table, and two 'Claude Fable 5' exploit cells match published Claude Mythos 5 results); Artificial Analysis chart lists eight proprietary models ahead, not five. Also: comparison-class definition, Lean case-study wording

No summary provided