COHERENCE - WikiHow: leaderboard
Metric: Exact-match accuracy (%) on the 2,076 WikiHow procedural 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 January 2027. 19 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 80.73 |
| 2 | GPT-5.4 (High) | 77.5 |
| 3 | Doubao-Seed-2.0-Pro-260215 | 76.59 |
| 4 | Kimi K2.5 | 75.67 |
| 5 | Claude Sonnet 4.6 (Thinking) | 74.18 |
| 6 | Qwen 3.5 122B A10B | 71.87 |
| 7 | Qwen 3.5 397B A17B | 69.85 |
| 8 | Qwen 3.5 35B A3B | 69.7 |
| 9 | GPT-5.2 | 68.5 |
| 10 | Qwen 3 VL 235B A22B | 64.45 |
| 11 | GLM-4.6V | 62.96 |
| 12 | Qwen 3.5 4B | 62.76 |
| 13 | Qwen 3 VL 4B Instruct | 58.57 |
| 14 | Step3 VL 10B | 57.47 |
| 15 | Qwen 3 VL 8B Instruct | 57.42 |
Interactive version: theaggregate.ai/benchmark?slug=coherence-wikihow · How It Works · Data refreshed daily, snapshot 2026-10-07.