Multiple-choice Question Answering on Text-only Adaptive Benchmark L5
65Pass@1 AccuracyQwen3-Omni
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen3-OmniModel Category=Omni models2025.12 | 65 | 0 | |
| Omni-AutoThinkModel Category=Omni models2025.12 | 48 | 53 | |
| AdaptThinkModel Category=Expert models2025.12 | 30 | 81 | |
| DeepSeek-R1-7BModel Category=Expert models2025.12 | 27 | 83 | |
| Qwen2.5-OmniModel Category=Omni models2025.12 | 7 | 23 |