List your company

Zero-Shot Generalist Tool Use on Toolathlon

36.4Success RateGemini-3-Pro
-1.4568.37218.228.028Mar 10, 2026

Evaluation results

MethodLinks
Gemini-3-ProCategory=Frontier (≫8B), Temperature=1, Top-p=1, Number of runs=32026.0336.4
Claude-4-SonnetCategory=Frontier (≫8B), Temperature=1, Top-p=1, Number of runs=32026.0329.9
DeepSeek-V3.2-ExpCategory=Frontier (≫8B), Temperature=1, Top-p=1, Number of runs=32026.0320.1
Kimi-K2-0905Category=Frontier (≫8B), Temperature=1, Top-p=1, Number of runs=32026.0313
Gemini-2.5-ProCategory=Frontier (≫8B), Temperature=1, Top-p=1, Number of runs=32026.0310.5
GPT-OSS-120BCategory=Frontier (≫8B), Temperature=1, Top-p=1, Number of runs=32026.039.8
DIVE-8B (RL)Category=Ours, Temperature=1, Top-p=1, Number of runs=32026.038.3
DIVE-8B (SFT)Category=Ours, Temperature=1, Top-p=1, Number of runs=32026.034.7
EnvScaler-8BCategory=8B Baselines, Temperature=1, Top-p=1, Number of runs=32026.032.2
Qwen3-8B (base)Category=Ours, Temperature=1, Top-p=1, Number of runs=32026.030.9
WebExplorer-8BCategory=8B Baselines, Temperature=1, Top-p=1, Number of runs=32026.030.3
SWE-Dev-8BCategory=8B Baselines, Temperature=1, Top-p=1, Number of runs=32026.030