ReactBench (Hallucination) - Alteration Tracing (CoT): leaderboard

Accuracy in percent on ReactBench (2026) questions about before-and-after image pairs, asking what changed, with a step-by-step reasoning prompt; an incorrect answer counts as a hallucination. Higher is better.

Metric: Accuracy (%) on Alteration Tracing (10,956 questions on 913 before-and-after image pairs) 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 February 2027. 13 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Qwen 3 VL 32B (Thinking)78.7#573
2Qwen 3.5 27B73.7#375
3ReactBench MiMo-VL-7B (checkpoint unspecified)73.5
4Qwen 3.5 9B63.9#828
5Qwen 3 VL 32B Instruct63.5#606
6InternVL3-8B60.7#1782
7Qwen 3 VL 4B53.4#1151
8Qwen 2.5 VL 72B Instruct50.5#841
9Qwen 2.5 VL 7B Instruct41.9#1902
10InternVL2.5-8B38.3#2370
11InternVL2.5-26B37.7#2156
12ReactBench LLaVA-OV-7B (checkpoint unspecified)34.7
13ReactBench LLaVA-v1.6-7B (checkpoint unspecified)34

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=reactbench-hallucination-alteration-tracing-cot · How It Works · Data refreshed daily, snapshot 2026-11-21.