Visual Commonsense Reasoning on VCR-EN-Easy
95.7EMInternVL2.5-78B
Evaluation results
| Method | Links | ||
|---|---|---|---|
| InternVL2.5-78B2024.12 | 95.7 | 94.5 | |
| InternVL2.5-38B2024.12 | 94.7 | 98.2 | |
| InternVL2.5-26B2024.12 | 94.4 | 98 | |
| InternVL2.5-4B2024.12 | 93.7 | 97.8 | |
| InternVL2.5-2B2024.12 | 93.2 | 97.6 | |
| InternVL2.5-8B2024.12 | 92.6 | 97.4 | |
| GPT-4o-202405132024.12 | 91.6 | 96.4 | |
| InternVL2.5-1B2024.12 | 91.5 | 97 | |
| Qwen2-VL-72B2024.12 | 91.3 | 94.6 | |
| Qwen2-VL-7B2024.12 | 89.7 | 93.8 | |
| InternVL2-40B2024.12 | 84.7 | 92.6 | |
| InternVL2-Llama3-76B2024.12 | 83.2 | 91.3 | |
| Qwen2-VL-2B2024.12 | 81.5 | — | |
| Cambrian-34B2024.12 | 79.7 | 89.3 | |
| InternVL2-26B2024.12 | 74.5 | 86.7 | |
| MiniCPM-V2.62024.12 | 73.9 | 85.7 | |
| Aquila-VL-2B2024.12 | 70 | — | |
| Claude-3.5-Sonnet2024.12 | 63.9 | 74.7 | |
| Gemini-1.5-Pro2024.12 | 62.7 | 77.7 | |
| Claude-3-Opus2024.12 | 62 | 77.7 | |
| GPT-4V2024.12 | 52 | 65.4 | |
| Phi-3.5-Vision-4B2024.12 | 39.3 | 60.4 | |
| InternVL2-8B2024.12 | 37.9 | 61.5 | |
| InternVL2-4B2024.12 | 33.7 | 61.1 | |
| InternVL2-2B2024.12 | 32.9 | 59.2 | |
| InternVL2-1B2024.12 | 21.5 | 48.4 | |
| InternVL-Chat-V1.52024.12 | 14.7 | 51.4 |