Sci-Rho - Worst-Case: leaderboard
Metric: Worst-case accuracy (%) over all seven languages on 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: the share of templates a model answers correctly on every one of its 10 generated variants, averaged over languages; higher is better. Source: arxiv.org. Saturation forecast: Around December 2026. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.4 (Non-reasoning) | 73.9 |
| 2 | Gemini 2.5 Pro | 62.5 |
| 3 | Gemini 2.5 Flash (Non-reasoning) | 56.6 |
| 4 | Qwen 3.5 27B (Non-reasoning) | 51.2 |
| 5 | Qwen 3.5 122B A10B (Non-reasoning) | 43.6 |
| 6 | Llama 4 Scout Instruct | 33.7 |
| 7 | Qwen 3 VL 8B Instruct | 29.5 |
| 8 | InternVL3.5-8B | 25.8 |
| 9 | Gemma 3 12B (IT) | 18.7 |
| 10 | Molmo2-8B | 15.5 |
| 11 | Gemma 3 4B (IT) | 7.2 |
| 12 | Qwen 3 VL 4B Instruct | 3.3 |
Interactive version: theaggregate.ai/benchmark?slug=sci-rho-worst-case · How It Works · Data refreshed daily, snapshot 2026-09-29.