Multimodal Reasoning on OlympiadBench
69.8Pass@1Gemini 1.5 Pro
Evaluation results
| Method | Links | |
|---|---|---|
| Gemini 1.5 Prothinking=true, decoding=greedy2025.05 | 69.8 | |
| Seed 1.5-VLthinking=true, decoding=greedy2025.05 | 65 | |
| Seed 1.5-VLthinking=false, decoding=greedy2025.05 | 60.4 | |
| Claude 3.7 Sonnetthinking=true, decoding=sampling2025.05 | 54.2 | |
| OpenAI o1thinking=true, decoding=greedy2025.05 | 48.5 | |
| Qwen 2.5-VL 72Bthinking=false, decoding=greedy2025.05 | 35.9 | |
| GPT-4othinking=false, decoding=greedy2025.05 | 25.9 |