HELM VHELM - Real World Reasoning: leaderboard
The "Real World Reasoning" subtask of HELM VHELM. Stanford HELM's VHELM project, the holistic evaluation of vision-language models: image-grounded question answering, OCR, chart and diagram reading, spatial reasoning, robustness, bias, fairness and toxicity scenarios scored under one ha...
Metric: Mean win rate. Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 22 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o (2024-05-13) | 86.9 | #326 |
| 2 | GPT-4o (2024-08-06) | 79.76 | #335 |
| 3 | Claude 3.5 Sonnet (20240620) | 75 | #298 |
| 4 | GPT-4o Mini (2024-07-18) | 72.62 | #561 |
| 5 | GPT-4 Turbo | 67.86 | #406 |
| 6 | Claude 3 Haiku (20240307) | 42.86 | #777 |
| 7 | Claude 3 Sonnet (20240229) | 42.86 | #681 |
| 8 | Claude 3 Opus (20240229) | 40.48 | #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-vhelm-real-world-reasoning · How It Works · Data refreshed daily, snapshot 2026-10-23.