Multimodal Multi-Turn Dialogue on MM-Snowball
68.25Turn-1 AccuracyGLM-4.6V-106B
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| GLM-4.6V-106BModel Category=Open-source2026.05 | 68.25 | 42.2 | 32.6 | 23.3 | 10.35 | 75.7 | 84.84 | 3.7 | |
| GPT-5Model Category=Proprietary2026.05 | 67.95 | 38.95 | 46.7 | 51.85 | 49.35 | 77.8 | 27.37 | 23.25 | |
| Seed-1.6Model Category=Proprietary2026.05 | 65.6 | 37.8 | 69.25 | 63.95 | 62.55 | 78.25 | 4.65 | 41.7 | |
| Kimi K2.5Model Category=Open-source2026.05 | 65.05 | 37.45 | 61.25 | 59.45 | 65.15 | 76.35 | -0.15 | 36.95 | |
| Gemini 2.5 FlashModel Category=Proprietary2026.05 | 62.25 | 34.95 | 50.95 | 24.25 | 10.6 | 74.4 | 82.97 | 3.25 | |
| Qwen3-VL-8BModel Category=Open-source2026.05 | 61.65 | 34 | 51.45 | 13.85 | 5.05 | 72.05 | 91.81 | 1.15 | |
| LLaVA-1.5-7BModel Category=Open-source2026.05 | 61.35 | 31.6 | 1.7 | 0.6 | 0.5 | 33.1 | 99.19 | 0 | |
| Qwen3-VL-32BModel Category=Open-source2026.05 | 59.95 | 33.3 | 56.95 | 26.7 | 9.35 | 69.35 | 84.4 | 4.15 | |
| Grok 4.1 FastModel Category=Proprietary2026.05 | 55.75 | 36.5 | 43.05 | 24.55 | 17.55 | 72.15 | 68.52 | 4.6 | |
| Qwen2.5-VL-7BModel Category=Open-source2026.05 | 51.3 | 26.95 | 32.75 | 8.6 | 3.15 | 46.9 | 93.86 | 0.15 | |
| Qwen2.5-VL-32BModel Category=Open-source2026.05 | 43.1 | 21.5 | 14.7 | 5.9 | 6.55 | 49.8 | 84.8 | 0.05 |