Text-Vision Reasoning on Text-Vision Benchmark L1 1.0 (test)
88Pass@1 AccuracyOmni-AutoThink
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Omni-AutoThinkModel Type=Omni models2025.12 | 88 | 52 | |
| Qwen3-OmniModel Type=Omni models2025.12 | 86 | 0 | |
| Qwen2.5-OmniModel Type=Omni models2025.12 | 85 | 0 | |
| Qwen2.5-VLModel Type=Expert models2025.12 | 82 | 24 | |
| R-4BModel Type=Expert models2025.12 | 20 | 55 |