Sci-Rho - Mathematics (English): leaderboard
Metric: Average-case accuracy (%) on the English mathematics templates of Sci-Rho: 606 expert-written executable templates per language (mathematics, physics, chemistry, biology and computer science, many drawn from olympiad problems), each rendered as 10 visually and numerically varied image-grounded instances; higher is better. Source: arxiv.org. Saturation forecast: Around December 2026. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.4 (Non-reasoning) | 90.3 |
| 2 | Gemini 2.5 Pro | 86.2 |
| 3 | Gemini 2.5 Flash (Non-reasoning) | 85.2 |
| 4 | Qwen 3.5 27B (Non-reasoning) | 85.1 |
| 5 | Qwen 3.5 122B A10B (Non-reasoning) | 82.1 |
| 6 | Llama 4 Scout Instruct | 71.4 |
| 7 | InternVL3.5-8B | 64.8 |
| 8 | Gemma 3 12B (IT) | 61.7 |
| 9 | Qwen 3 VL 8B Instruct | 61.5 |
| 10 | Molmo2-8B | 46.5 |
| 11 | Gemma 3 4B (IT) | 46.2 |
| 12 | Qwen 3 VL 4B Instruct | 31.7 |
Interactive version: theaggregate.ai/benchmark?slug=sci-rho-mathematics-english · How It Works · Data refreshed daily, snapshot 2026-09-29.