Multimodal Tool-Use on MMSearch-Plus
33.53Avg@4Kimi-K2.5
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Kimi-K2.5Evaluation Protocol=Vanilla Tools2026.05 | 33.53 | 48.34 | |
| Qwen3.5-35B-A3BEvaluation Protocol=w/ ACE-SKILL2026.05 | 28.55 | 43.13 | |
| Qwen3.5-PlusEvaluation Protocol=Vanilla Tools2026.05 | 27.84 | 46.45 | |
| Qwen3.5-35B-A3BEvaluation Protocol=w/ XSkill2026.05 | 23.93 | 41.23 | |
| Gemini-2.5-ProEvaluation Protocol=Vanilla Tools2026.05 | 21.56 | 35.55 | |
| GLM-4.6VEvaluation Protocol=Vanilla Tools2026.05 | 17.42 | 34.6 | |
| Qwen3.5-35B-A3BEvaluation Protocol=Vanilla Tools2026.05 | 16.59 | 28.44 | |
| o4-miniEvaluation Protocol=Vanilla Tools2026.05 | 15.88 | 21.8 | |
| GPT-5-miniEvaluation Protocol=Vanilla Tools2026.05 | 14.22 | 20.38 | |
| Kimi-K2.5Evaluation Protocol=Native Direct Answer2026.05 | 10.78 | 18.48 | |
| Gemini-2.5-ProEvaluation Protocol=Native Direct Answer2026.05 | 10.43 | 19.43 | |
| Qwen3.5-PlusEvaluation Protocol=Native Direct Answer2026.05 | 6.64 | 10.43 | |
| o4-miniEvaluation Protocol=Native Direct Answer2026.05 | 5.92 | 12.32 | |
| GLM-4.6VEvaluation Protocol=Native Direct Answer2026.05 | 3.67 | 8.06 | |
| Qwen3.5-35B-A3BEvaluation Protocol=Native Direct Answer2026.05 | 3.67 | 7.58 | |
| GPT-5-miniEvaluation Protocol=Native Direct Answer2026.05 | 3.08 | 6.64 |