OpenCompass Multimodal - HallusionBench: leaderboard
OpenCompass multimodal score on HallusionBench visual hallucination and reasoning checks.
Metric: Score (%). Source: rank.opencompass.org.cn. 105 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | BlueLM-3.5-Nano-3B | 70.5 | |
| 2 | SenseNova-V6-5-Pro | 66.7 | |
| 3 | TeleMM-2.0 | 65.7 | |
| 4 | Ovis2.5-9B | 65.3 | |
| 5 | TeleMM-2.0-Thinking | 65.2 | |
| 6 | GPT-5-20250807 | 65.2 | |
| 7 | MiniCPM-o-4_5 | 65 | |
| 8 | Gemini-2.5-Pro | 64.1 | |
| 9 | JT-VL-Chat-Mini | 64.1 | |
| 10 | JT3.5 | 63.8 |
Interactive version: theaggregate.ai/benchmark?slug=opencompass-multimodal-hallusionbench · How It Works · Data refreshed daily, snapshot 2026-10-20.