Agentic Tool Use on Tool-Decathlon
46.3ScoreGPT-5.2 (xhigh)
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-5.2 (xhigh)2026.02 | 46.3 | |
| Claude Opus 4.52026.02 | 43.5 | |
| GLM-52026.02 | 39.2 | |
| Gemini 3 Pro2026.02 | 36.4 | |
| DeepSeek-V3.22026.02 | 35.2 | |
| Kimi K2.52026.02 | 27.8 | |
| GLM-4.72026.02 | 23.8 |