Artificial Analysis Omniscience Accuracy
General QA · 2025-11-16
Artificial Analysis Omniscience Accuracy measures how often a model answers the evaluation's 6,000 fact-checking or knowledge questions correctly. It is paired with a hallucination-rate metric.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Opus 5.5 | 66.2 |
| GPT-6 Astra | 62.6 |
| GPT-6.1 Sol | 62.1 |
| Claude Opus 5 | 60.9 |
| GPT-5.6 Sol | 59.4 |
| GPT-5.5 | 58.0 |
| MiniMax M3.1 Flash | 56.2 |
| Gemini 3 Pro Preview | 55.8 |
| Gemini 3.7 Flash | 55.3 |
| Gemini 3.1 Pro Preview | 54.9 |
| Gemini 3.8 Flash | 54.6 |
| GPT-6 Sol | 54.5 |
| Claude Sonnet 5.5 | 54.0 |
| Gemini 3 Flash Preview | 53.4 |
| GPT-5.3-Codex | 52.9 |
| Muse Spark 1.1 | 52.0 |
| Grok 4.5 | 51.5 |
| Grok Build 0.1 | 51.5 |
| Gemini 3.5 Flash | 51.4 |
| Kimi K3.1 Pro | 51.0 |
| GPT-5.4 | 50.9 |
| Gemini 3.6 Flash | 50.0 |
| GPT-6.1 Terra | 50.0 |
| Gemini 4 Argon | 49.9 |
| Grok 4.7 | 49.7 |
| Muse Spark | 49.6 |
| JT-4.1 Flash 236B A21B | 49.4 |
| DeepSeek V4 Pro 0813 | 49.1 |
| Opus 4.7 | 48.9 |
| Opus 4.8 | 48.8 |
| Grok 4.6 | 48.2 |
| Kimi K3 | 47.6 |
| Opus 4.6 | 47.0 |
| GPT-5.6 Terra | 46.8 |
| Opus 4.5 | 46.6 |
| DeepSeek V4.1 Flash | 46.4 |
| GPT-5.5 Instant | 46.3 |
| Muse Spark 1.2 | 45.4 |
| GPT-5.5 Instant (2026-06-25 hosted snapshot) | 44.4 |
| GPT-5.2 | 44.3 |