Multimodal Reasoning on SPD-Faith Bench Hard 1.0
10.9Contradiction RateMiniCPM-V-2.6
Evaluation Results
| Method | Links | |
|---|---|---|
| MiniCPM-V-2.6Model Type=Open-Source2026.02 | 10.9 | |
| Qwen2.5-VL-72BModel Type=Open-Source, Parameters=72B2026.02 | 11.4 | |
| InternVL2.5-38BModel Type=Open-Source, Parameters=38B2026.02 | 11.8 | |
| Qwen2.5-VL-7BModel Type=Open-Source, Parameters=7B2026.02 | 12.2 | |
| InternVL2.5-8BModel Type=Open-Source, Parameters=8B2026.02 | 12.5 | |
| Qwen3-VL-32BModel Type=Open-Source, Parameters=32B2026.02 | 12.9 | |
| Qwen3-VL-235B-A22BModel Type=Open-Source, Parameters=235B-A22B2026.02 | 17 | |
| GLM-4.5VModel Type=Open-Source2026.02 | 19.5 | |
| GPT-4oModel Type=Proprietary2026.02 | 22.8 | |
| Gemini-2.5-ProModel Type=Proprietary2026.02 | 27.2 | |
| DeepSeek-VL2Model Type=Open-Source2026.02 | 32.5 | |
| Claude-4.5-HaikuModel Type=Proprietary2026.02 | 37.5 |