Interactive Agent Task on TRIP-Bench Easy
71Loose Success CountDeepSeek-V3.2
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DeepSeek-V3.2thinking mode=w/ thinking2026.02 | 71 | 31 | |
| GPT-5.2thinking mode=w/ thinking2026.02 | 66 | 49 | |
| Claude-Sonnet-4.5thinking mode=w/ thinking2026.02 | 58 | 27 | |
| Gemini-3-Prothinking mode=w/o thinking2026.02 | 44 | 12 | |
| Gemini-3-Flashthinking mode=w/ thinking2026.02 | 44 | 25 | |
| Gemini-3-Prothinking mode=w/ thinking2026.02 | 42 | 11 | |
| DeepSeek-V3.2thinking mode=w/o thinking2026.02 | 39 | 5 | |
| GLM-4.7thinking mode=w/ thinking2026.02 | 38 | 16 | |
| Claude-Sonnet-4.5thinking mode=w/o thinking2026.02 | 36 | 7 | |
| Gemini-3-Flashthinking mode=w/o thinking2026.02 | 36 | 22 | |
| Kimi-K2-Thinkingthinking mode=w/ thinking2026.02 | 35 | 5 | |
| GLM-4.7thinking mode=w/o thinking2026.02 | 34 | 0 | |
| GPT-5.2thinking mode=w/o thinking2026.02 | 24 | 2 | |
| Qwen3-235B-A22B-Instruct-2507thinking mode=w/o thinking2026.02 | 16 | 2 | |
| Kimi-K2-0905-Previewthinking mode=w/o thinking2026.02 | 13 | 0 | |
| Qwen3-235B-A22B-Thinking-2507thinking mode=w/ thinking2026.02 | 0 | 0 |