Benchgen

TAU-bench Airline — Results

RankModelScore
1claude-sonnet-4-50.7
2minimax-m1-80k0.62
3glm-4-5-air0.608
4glm-4-50.604
5claude-sonnet-40.6
6minimax-m1-40k0.6
7qwen3-coder-480b-a35b0.6
8claude-opus-40.596
9claude-opus-4-10.56
10gpt-4-50.5
11o10.5
12gpt-4-10.494
13o4-mini0.492
14qwen3-next-80b-a3b-thinking0.49
15claude-3-5-sonnet0.46
16qwen3-235b-a22b-thinking-25070.46
17qwen3-next-80b-a3b-instruct0.44
18gpt-4o0.428
19gpt-4-1-mini0.36
20o3-mini0.324
21claude-3-5-haiku0.228

TAU-bench Airline — Phases

1 phaseActive

Phase 1·

TAU-bench Airline

Aug 4, 2026
Current