HELM Image2Struct - Image2Webpage: leaderboard
The "Image2Webpage" subtask of HELM Image2Struct. Stanford HELM's Image2Struct project: a model is shown a rendered image — a webpage, a LaTeX plot or equation, a music score — and must emit the source that reproduces it, scored on how closely the re-rendered output matches the original.
Metric: EMS. Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 20 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o (2024-08-06) | 73.51 | #335 |
| 2 | Claude 3.5 Sonnet (20240620) | 72.41 | #298 |
| 3 | Claude 3.5 Sonnet (20241022) | 71.22 | #291 |
| 4 | GPT-4o (2024-05-13) | 71.03 | #326 |
| 5 | GPT-4o Mini (2024-07-18) | 69.59 | #561 |
| 6 | Gemini 1.5 Pro (002) | 68.32 | #350 |
| 7 | Gemini 1.5 Flash (002) | 67.37 | #585 |
| 8 | Claude 3 Sonnet (20240229) | 64.19 | #681 |
| 9 | Claude 3 Opus (20240229) | 37.8 | #413 |
No result here: #4 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-image2struct-image2webpage · How It Works · Data refreshed daily, snapshot 2026-10-23.