Science Reasoning on GPQA Diamond (Metric: GPQA Diamond)
54.6GPQA DiamondQwen3-32B
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3-32BTarget LLM=Public Models2026.05 | 54.6 | |
| Qwen3-4BTarget LLM=Public Models2026.05 | 41.7 | |
| DPA-Qwen3-32BTarget LLM=Qwen3-32B2026.05 | 35.4 | |
| LLaVA-NeXT-Qwen3-32BTarget LLM=Qwen3-32B2026.05 | 35.3 | |
| DPA-Qwen3-4BTarget LLM=Qwen3-4B2026.05 | 33.3 | |
| LLaVA-NeXT-Qwen3-4BTarget LLM=Qwen3-4B2026.05 | 32.8 | |
| DPA-LLaMA-3.2-3BTarget LLM=LLaMA-3.2-3B2026.05 | 25.3 | |
| LLaVA-NeXT-LLaMA-3.2-3BTarget LLM=LLaMA-3.2-3B2026.05 | 17.2 | |
| Qwen2-VL-2BTarget LLM=Public Models2026.05 | 15.7 | |
| Idefics2-8BTarget LLM=Public Models2026.05 | 15.2 | |
| LLaVA-1.5-7BTarget LLM=Public Models2026.05 | 12.6 | |
| LLaVA-NeXT-7BTarget LLM=Public Models2026.05 | 12.1 | |
| Cambrian-1-8BTarget LLM=Public Models2026.05 | 0 |