Agent (Tool Use) on OmniGAIA
68.9ScoreGemini-3.1 Pro
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini-3.1 Prothinking prompt=without, answer formatting=without, evaluation judge=DeepSeek-V3.2-Thinking2026.04 | 68.9 | |
| Qwen3.5-Omni-Plusthinking prompt=without, answer formatting=without, evaluation judge=DeepSeek-V3.2-Thinking2026.04 | 57.2 | |
| Qwen3.5-Omni-Flashthinking prompt=without, answer formatting=without, evaluation judge=DeepSeek-V3.2-Thinking2026.04 | 33.9 |