HarmVideoBench: leaderboard
Metric: Macro Avg. (self-reported). Source: benchmarklist.com. 21 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 82.1 |
| 2 | Gemini 3.1 Flash | 81.3 |
| 3 | GPT-5.5 | 80.9 |
| 4 | Claude Opus 4.7 | 80.3 |
| 5 | Qwen 3 VL 235B A22B (Thinking) | 80.1 |
| 6 | Claude Sonnet 4.6 | 79.9 |
| 7 | Qwen 3 VL 235B A22B Instruct | 79.7 |
| 8 | Qwen 3 VL 32B (Thinking) | 79.6 |
| 9 | Qwen 3 VL 32B Instruct | 79 |
| 10 | Qwen 3 VL 30B A3B (Thinking) | 78.9 |
| 11 | Qwen 3 VL 8B (Thinking) | 76.7 |
| 12 | Qwen 3 VL 30B A3B Instruct | 76.3 |
| 13 | Kimi K2.5 | 71.2 |
| 14 | LFM2.5-VL-1.6B | 66.8 |
| 15 | GLM-4.5V | 65.5 |
Interactive version: theaggregate.ai/benchmark?slug=harmvideobench · How It Works · Data refreshed daily, snapshot 2026-09-05.