MoHallBench - Semantic Similarity Hallucination: leaderboard
Metric: Q-PairAcc (%; share of adversarial pairs in which the model answers both the positive and the negative (reversed) yes/no query correctly; 32 uniformly sampled frames, temperature 0; semantically related but physically distinct actions). Source: arxiv.org. Saturation forecast: Around December 2026. 10 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 VL 8B Instruct | 39.3 |
| 2 | Molmo2-8B | 27 |
| 3 | InternVL3-8B | 25.8 |
| 4 | Qwen 2.5 VL 7B Instruct | 20.7 |
| 5 | Qwen 2 VL 7B Instruct | 16.4 |
Interactive version: theaggregate.ai/benchmark?slug=mohallbench-semantic-similarity-hallucination · How It Works · Data refreshed daily, snapshot 2026-09-29.