ReactBench (Hallucination) - Counterfactual Attribute (CoT): leaderboard
Accuracy in percent on ReactBench (2026) questions about images whose object attribute was edited against expectation, with a step-by-step reasoning prompt; an incorrect answer counts as a hallucination. Higher is better.
Metric: Accuracy (%) on Counterfactual Attribute (7,668 questions on 639 images with an edited attribute) with a step-by-step reasoning prompt, hallucination-inducing questions (multiple choice, short answer, yes/no) about images edited by Qwen-Image-Edit-2511; an incorrect answer counts as a hallucination; temperature 0.2, at most 1,024 tokens, LLM answer matching; higher is better. Source: arxiv.org. Saturation forecast: Around May 2027. 13 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3 VL 32B Instruct | 63.4 | #607 |
| 2 | InternVL3-8B | 60.9 | #1795 |
| 3 | Qwen 2.5 VL 72B Instruct | 60.7 | #849 |
| 4 | Qwen 3 VL 4B | 60.6 | #1163 |
| 5 | Qwen 3 VL 32B (Thinking) | 60.5 | #577 |
| 6 | InternVL2.5-26B | 59.7 | #2174 |
| 7 | Qwen 3.5 27B | 59.5 | #380 |
| 8 | Qwen 3.5 9B | 57.6 | #838 |
| 9 | Qwen 2.5 VL 7B Instruct | 55.2 | #1919 |
| 10 | ReactBench MiMo-VL-7B (checkpoint unspecified) | 50.8 | |
| 11 | InternVL2.5-8B | 46.9 | #2393 |
| 12 | ReactBench LLaVA-OV-7B (checkpoint unspecified) | 44.8 | |
| 13 | ReactBench LLaVA-v1.6-7B (checkpoint unspecified) | 37.5 |
No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=reactbench-hallucination-counterfactual-attribute-cot · How It Works · Data refreshed daily, snapshot 2026-11-22.