Text-Vision Reasoning on Text-Vision Benchmark L3 1.0 (test)
59Pass@1 AccuracyOmni-AutoThink
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Omni-AutoThinkModel Type=Omni models2025.12 | 59 | 69 | |
| Qwen2.5-OmniModel Type=Omni models2025.12 | 58 | 0 | |
| Qwen3-OmniModel Type=Omni models2025.12 | 52 | 0 | |
| Qwen2.5-VLModel Type=Expert models2025.12 | 48 | 37 | |
| R-4BModel Type=Expert models2025.12 | 21 | 56 |