DashBench - Q5: leaderboard
DashBench evaluates video understanding on 1,709 driving clips. Five explanation prompts and binary event detection are scored separately; mean agreement excludes the binary task. The 823-clip corner-case-window subset and training-free decoder variants retain their own conditions. Q5 asks for definition-guided corner-case identification and description. This board reports lLM-judged answer accuracy (%). Higher is better.
Metric: LLM-judged answer accuracy (%). Source: arxiv.org. Saturation forecast: Around February 2027. 22 models tracked. Updated 14 Nov 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | GPT-5 | 39.15 | #167 | |
| 2 | Gemini 3 Flash | 34.99 | #163 | |
| 3 | Qwen 3 VL 235B A22B | 21.71 | #484 | |
| 4 | Qwen 3 VL 8B | 21.53 | #899 | |
| 5 | Ovis2.5-9B | 20.55 | #1236 | |
| 6 | Qwen 3 VL 32B | 18.31 | #609 | |
| 7 | InternVL3.5-241B | 17.61 | #625 | |
| 8 | InternVL3.5-8B | Evaluation run: DashBench TRiCD | 14.69 | #1558 |
| 9 | MiniCPM-V-4.5 | 13.53 | #1390 | |
| 10 | InternVL3.5-38B | 12.93 | #986 | |
| 11 | Tarsier2-7B | 12.35 | #844 | |
| 12 | InternVideo2.5-Chat-8B | 10.12 | #2017 | |
| 13 | GLM-4.1V-9B (Thinking) | 9.91 | ||
| 14 | InternLM-XComposer2.5 | 8.54 | #3981 | |
| 15 | LlaVA-OV-RS (7B) | 7.49 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=dashbench-q5 · How It Works · Data refreshed daily, snapshot 2026-11-16.