LiveBench Overall ↗2.0 · 2026-10-14 (dataset 2026-09-02)
40.83%Effort: max · single run · equal mean of 7 suites
Effort: max · single run · equal mean of 7 suites
Effort: max · single run
Effort: max · single run
Effort: max · single run
Effort: max · single run
Effort: max · single run
Effort: max · single run
Effort: max · single run
120B/12B-active Mamba-2 hybrid with a vision encoder; native speculative decoding and switchable thinking.