COHERENCE - StoryBird: 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 on the StoryBird story documents, in percent; higher is better.

Metric: Exact-match accuracy (%) on the 1,398 StoryBird story documents; 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

#ModelSettingScoreOverall rank
1Gemini 3.1 Pro (Preview)63.59#99
2GPT-5.4High57.37#135
3Qwen 3.5 397B A17B55.58#264
4Claude Sonnet 4.6Thinking52.72#173
5Doubao-Seed-2.0-Pro-26021552.22#158
6Qwen 3.5 122B A10B49.93#346
7Qwen 3.5 35B A3B48.21#492
8Kimi K2.546.21#266
9Doubao-Seed-2.0-Lite-26021544.35#262
10Doubao-Seed-2.0-Mini-26021537.98
11GPT-5.237.34#185
12Qwen 3.5 4B36.41#1241
13Qwen 3 VL 235B A22B32.12#485
14GLM-4.6V29.18#710
15Intern-S1-Pro28.9#585

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

Interactive version: theaggregate.ai/benchmark?slug=coherence-storybird · How It Works · Data refreshed daily, snapshot 2026-11-21.