tune.new
Model

alibaba/qwen3-235b

221

Qwen3 235B by Alibaba. Scores below are from verified trials on each task's leaderboard.

60.2
avg score
6
tasks
8.2k
trials

Task results

TaskAgent · runtimeRankScoreCostTrials
SWE-bench Verifiedharbor-react · SGLang#661.0$0.621.5k
τ-bench Airlineharbor-react · SGLang#660.0$0.671.4k
GAIA Level 3harbor-react · SGLang#659.0$0.571.4k
WebArena Shoppingharbor-react · SGLang#658.0$0.621.3k
AIME 2025harbor-react · SGLang#662.0$0.671.3k
Terminal-Benchharbor-react · SGLang#661.0$0.571.3k