COHERENCE: leaderboard
COHERENCE (2026) removes the images from 6,161 interleaved image-text documents (WikiHow, StoryBird, cooking recipes, science) and asks the model to put each candidate image back in its place. This board is the exact-match accuracy over all documents, the whole assignment right, in percent; higher is better.
Metric: Exact-match accuracy (%): the whole image-to-placeholder assignment is right, over all 6,161 interleaved image-text documents whose images were removed: the model assigns each candidate image to its placeholder; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 19 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 71.82 | #99 | |
| 2 | GPT-5.4 | High | 71.29 | #135 |
| 3 | Claude Sonnet 4.6 | Thinking | 65.72 | #173 |
| 4 | Qwen 3.5 397B A17B | 64.81 | #264 | |
| 5 | Doubao-Seed-2.0-Pro-260215 | 62.1 | #158 | |
| 6 | Kimi K2.5 | 60.19 | #266 | |
| 7 | Qwen 3.5 122B A10B | 59.57 | #346 | |
| 8 | Qwen 3.5 35B A3B | 55.82 | #492 | |
| 9 | Doubao-Seed-2.0-Lite-260215 | 55.79 | #262 | |
| 10 | GPT-5.2 | 52.46 | #185 | |
| 11 | Doubao-Seed-2.0-Mini-260215 | 50.24 | ||
| 12 | Intern-S1-Pro | 47.25 | #585 | |
| 13 | Qwen 3 VL 235B A22B | 46.32 | #485 | |
| 14 | Qwen 3 VL 30B A3B | 46.24 | #784 | |
| 15 | Qwen 3.5 4B | 45.64 | #1241 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=coherence · How It Works · Data refreshed daily, snapshot 2026-11-21.