Rabdos Vizi Bench
For reasoning over scientific and mathematical visualizations.
Sample visualization
Leaderboard
27 problems · Updated Jul 31, 2026
| Rank | Provider | Model | Reasoning effort | Graded score |
|---|---|---|---|---|
| 1 | Anthropic | Claude Fable 5 | max | 33.3% |
| 2 | OpenAI | GPT-5.6 Sol | max | 29.6% |
| 3 | Gemini 3.6 Flash | high | 18.5% | |
| 4 | xAI | Grok 4.5 | xhigh | 14.8% |
| 5 | Anthropic | Claude Opus 4.8 | max | 14.8% |
| 6 | Moonshot AI | Kimi K3 | max | 11.1% |
| 7 | Meta | Muse Spark 1.1 | xhigh | 3.7% |
Interested in evaluating your model on Rabdos Vizi Bench? Get in touch →