ZeroBench Main Questions pass^5
Multimodal · 2025-02-13
ZeroBench pass^5 is the main-question consistency metric on ZeroBench, a visual reasoning benchmark for large multimodal models. It is computed from five samples and reported alongside pass@5 on the official leaderboard.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-6 Astra | 35.0 |
| Claude Opus 5.5 | 25.0 |
| GPT-6 Sol | 17.0 |
| GPT-5.6 Sol | 13.0 |
| Claude Opus 5 | 11.0 |
| GPT-5.5 | 11.0 |
| GPT-5.6 Terra | 10.0 |
| Grok 4.5 | 9.0 |
| Claude Fable 5 | 9.0 |
| GPT-5.4 | 9.0 |
| GPT-5.6 Luna | 8.0 |
| Gemini 3.6 Flash | 8.0 |
| Gemini 3.1 Pro Preview | 8.0 |
| GPT-6 Luna | 7.0 |
| Grok 4.6 | 7.0 |
| Opus 4.8 | 6.0 |
| GPT-5.4 Mini | 6.0 |
| GPT-5.2 | 6.0 |
| Sonnet 5 | 5.0 |
| Gemini 3.5 Flash | 5.0 |
| Opus 4.7 | 5.0 |
| Gemini 3 Pro Preview | 5.0 |
| GPT-5 Mini | 4.0 |
| Sonnet 4.6 | 2.0 |
| Opus 4.6 | 2.0 |
| Gemini 3 Flash Preview | 2.0 |
| Grok 4.3 | 1.0 |
| Gemini 3.1 Flash-Lite Preview | 1.0 |
| Opus 4.5 | 1.0 |
| Sonnet 4.5 | 1.0 |
| Opus 4.1 | 1.0 |
| Opus 4 | 1.0 |
| Sonnet 4 | 1.0 |
| Llama 4 Scout Instruct | 1.0 |
| Gemini 2.5 Pro Experimental 03-25 | 1.0 |
| Gemini 2.5 Flash Preview 04-17 | 1.0 |
| Claude 3.7 Sonnet | 1.0 |
| Gemini 3.5 Flash-Lite | 0.0 |
| GPT-5.1 | 0.0 |
| GPT-5 | 0.0 |