DORA - Multimodal Report Synthesis: leaderboard
Metric: Final-answer accuracy (%; ReAct agent with tools). Source: github.com. Saturation forecast: Around January 2028. 13 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.5 397B A17B | 49.17 |
| 2 | Claude Sonnet 4.6 | 45.53 |
| 3 | Gemini 3 Flash | 45.31 |
| 4 | Grok 4.1 Fast | 44.52 |
| 5 | MiMo-V2-Pro | 44.26 |
| 6 | Gemma 4 31B | 42.03 |
| 7 | Step 3.5 Flash | 41.58 |
| 8 | DeepSeek V3.2 | 41.01 |
| 9 | MiniMax-M2.7 | 40.62 |
| 10 | GPT-5.4 | 29.11 |
| 11 | GPT-OSS-120B | 28.84 |
| 12 | GPT-5.4 Nano | 27.37 |
| 13 | Qwen 3.5 35B A3B | 21.3 |
Interactive version: theaggregate.ai/benchmark?slug=dora-multimodal-report-synthesis · How It Works · Data refreshed daily, snapshot 2026-10-02.