Multimodal Reasoning on VisuLogic
35Pass@1Seed 1.5-VL
Evaluation Results
| Method | Links | |
|---|---|---|
| Seed 1.5-VLthinking=true, decoding=greedy2025.05 | 35 | |
| Seed 1.5-VLthinking=false, decoding=greedy2025.05 | 33 | |
| Gemini 1.5 Prothinking=true, decoding=greedy2025.05 | 31 | |
| OpenAI o1thinking=true, decoding=greedy2025.05 | 29 | |
| MOSS-Video-PreviewSFT Mode=offline2026.06 | 28.7 | |
| MOSS-Video-PreviewSFT Mode=real-time2026.06 | 28.6 | |
| Qwen 2.5-VL 72Bthinking=false, decoding=greedy2025.05 | 28 | |
| Uni-OPDDistillation Setting=Single-Teacher, N-Shot=02026.05 | 27.6 | |
| Qwen3-VL-4B-Instruct-Logic-RL TeacherN-Shot=02026.05 | 27.4 | |
| OPDDistillation Setting=Multi-Teacher, N-Shot=02026.05 | 27.2 | |
| Uni-OPDDistillation Setting=Multi-Teacher, N-Shot=02026.05 | 27 | |
| LLaVA-OVScale=1.5-8B2026.06 | 27 | |
| OPDDistillation Setting=Single-Teacher, N-Shot=02026.05 | 26.8 | |
| UnicornText-only=true, Pretrain Scale=1.2M2025.03 | 26.8 | |
| GPT-4othinking=false, decoding=greedy2025.05 | 26.3 | |
| Qwen2.5-VLScale=7B2026.06 | 25.9 | |
| BunnyText-only=false, Pretrain Scale=1.2M2025.03 | 25.3 | |
| Claude 3.7 Sonnetthinking=true, decoding=sampling2025.05 | 24.8 | |
| Unicorn*Text-only=true, Pretrain Scale=558K2025.03 | 24.3 | |
| LLaVA-1.5Text-only=false, Pretrain Scale=558K2025.03 | 22.9 | |
| Qwen3-VL-2B-Instruct StudentN-Shot=02026.05 | 6.4 |