LLM Stats (Hallusion Bench): leaderboard
Metric: Score (%). Source: llm-stats.com. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.5 27B | 70 |
| 2 | Qwen 3.6 35B A3B | 69.8 |
| 3 | Qwen 3.5 35B A3B | 67.9 |
| 4 | Qwen 3.5 122B A10B | 67.6 |
| 5 | Qwen 3 VL 32B (Thinking) | 67.4 |
| 6 | Qwen 3 VL 235B A22B (Thinking) | 66.7 |
| 7 | Qwen 3 VL 30B A3B (Thinking) | 66 |
| 8 | Qwen 3 VL 8B (Thinking) | 65.4 |
| 9 | Qwen 3 VL 4B (Thinking) | 64.1 |
| 10 | Qwen 3 VL 32B Instruct | 63.8 |
| 11 | Qwen 3 VL 235B A22B Instruct | 63.2 |
| 12 | Qwen 3 VL 30B A3B Instruct | 61.5 |
| 13 | Qwen 3 VL 8B Instruct | 61.1 |
| 14 | Qwen 3 VL 4B Instruct | 57.6 |
| 15 | Qwen 2.5 VL 7B Instruct | 52.9 |
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-hallusion-bench · How It Works · Data refreshed daily, snapshot 2026-09-05.