Multimodal long-horizon complex reasoning on MM-BrowseComp
41.1ScoreClaude-4.6-Sonnet-Thinking
Evaluation Results
| Method | Links | |
|---|---|---|
| Claude-4.6-Sonnet-ThinkingSize=-, Train=-2026.06 | 41.1 | |
| DeepSeek-V3.2Size=671B, Train=-2026.06 | 31.7 | |
| Doubao-2.0-ProSize=-, Train=-2026.06 | 27.2 | |
| Qwen3.5-397BSize=397B, Train=-2026.06 | 27.2 | |
| REDSearcher-MM-SFTSize=30B, Train=Mid/SFT2026.06 | 25.3 | |
| GLM-5Size=744B, Train=-2026.06 | 22.8 | |
| GPT-5.2Size=-, Train=-2026.06 | 22.3 | |
| Kimi-K2.5Size=1T, Train=-2026.06 | 20.5 | |
| MiniMax-M2.7Size=230B, Train=-2026.06 | 20.1 | |
| S1-DeepResearchSize=32B, Train=SFT2026.06 | 19.2 | |
| Gemini-3.1-Pro-PreviewSize=-, Train=-2026.06 | 11.6 | |
| Qwen3-32BSize=32B, Train=-2026.06 | 7.1 | |
| Qwen3-235BSize=235B, Train=-2026.06 | 6.7 |