context.vn
Menu

tau3 Bench

17 models evaluated

#ModelProviderTypeScore
1Mercury 2.5InceptionProprietaryτ³-bench Telecom
2Mistral Medium 3.5 128BMistralOpen weightsτ³-bench Telecom
4Nemotron 3 UltraNVIDIA · Open weightτ³-bench published average70.9%
5Qwen3.6 PlusAlibaba · Closedτ³-bench published setup70.7%
6GLM-5.1Z.AI · Open weightτ³-bench published setup70.6%
7Claude Opus 4.5Anthropic · Closedτ³-bench published setup70.2%
8Qwen3.5 397BAlibaba · Open weightτ³-bench published setup68.4%
9Qwen3.6-35B-A3BAlibaba · Open weightτ³-bench published setup67.2%
10Pokee-Isaac 28BPokee AI · Closedτ³-bench Telecom66.2%
11Kimi K2.5Moonshot AI · Open weightτ³-bench published setup65.7%
12GLM-5Z.AI · Open weightτ³-bench published setup65.6%
13Granite 4.2 30BIBM · Open weightτ³-bench published setup62.0%
14Granite 4.2 8BIBM · Open weightτ³-bench published average58.1%
15Granite 4.2 3BIBM · Open weightτ³-bench published setup45.8%
17Nemotron 3.5 Lightning 30B A3B NVFP4NVIDIA · Open weightτ³-bench published setup9.5%
18Mercury 2Inception · Closedτ³-bench published setup9.0%
19LFM2.5-2.6BLiquidAI · Open weightτ³-bench published setup5.7%