Visual Question Answering on SeedBench (Sub-category Scores)
78.9Average ScorePerceptionDLM-Base
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| PerceptionDLM-BaseModel Size=8B2026.06 | 78.9 | — | — | — | — | — | — | — | — | — | |
| InternVL3Model Size=8B2026.06 | 77.1 | — | — | — | — | — | — | — | — | — | |
| Qwen2.5-VLModel Size=7B2026.06 | 77 | — | — | — | — | — | — | — | — | — | |
| Dream-VLModel Size=7B2026.06 | 76.4 | — | — | — | — | — | — | — | — | — | |
| SDAR-VLModel Size=8B2026.06 | 75.5 | — | — | — | — | — | — | — | — | — | |
| LLaDA-VModel Size=8B2026.06 | 74.8 | — | — | — | — | — | — | — | — | — | |
| Qwen2.5VL-3B + CompART2024.12 | 65 | 66.8 | 74.6 | 56.9 | 68.7 | 47.2 | 58.8 | 64.7 | 78.3 | 69.2 | |
| MMaDAModel Size=8B2026.06 | 64.2 | — | — | — | — | — | — | — | — | — | |
| Qwen2.5VL-3B2024.12 | 62.8 | 64.9 | 69.5 | 52.3 | 67.4 | 44.1 | 59.8 | 62.8 | 78.3 | 66.2 | |
| SmolVLM-2.2B2024.12 | 62.6 | 64 | 73.4 | 56 | 63.4 | 41.9 | 54.6 | 64.1 | 76.5 | 69.6 | |
| InternVL3.5-4B2024.12 | 62.2 | 62.4 | 67.8 | 53.6 | 68.8 | 47.3 | 60.8 | 63.4 | 70.6 | 65.5 | |
| LLaVA1.5-7B2024.12 | 58 | 61.7 | 66.2 | 49.1 | 58.5 | 38.5 | 52.6 | 64.4 | 67.1 | 63.8 | |
| Gemma3-4B2024.12 | 53.2 | 56.7 | 59.9 | 42.3 | 52.6 | 39.1 | 52.6 | 54.1 | 63.5 | 57.8 |