AA-Omniscience Index
Indexes · 2025-11-16
Factual recall and hallucination benchmark across 6,000 questions in economically relevant domains; higher scores mean more correct answers and fewer hallucinations.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Opus 5.5 | 46.4 |
| GPT-6 Astra | 43.7 |
| Gemini 4 Argon | 42.4 |
| GPT-6.1 Sol | 41.5 |
| Claude Opus 5 | 37.1 |
| Claude Sonnet 5.5 | 32.3 |
| Grok 4.7 | 32.0 |
| Gemini 3.1 Pro Preview | 31.9 |
| Grok 4.6 | 30.5 |
| Muse Wonder Full | 30.1 |
| GPT-6.1 Terra | 29.8 |
| Gemini 3.8 Flash | 29.6 |
| Kimi K3.1 Pro | 29.5 |
| Opus 4.8 | 28.8 |
| Muse Spark 1.1 | 28.1 |
| Opus 4.7 | 27.3 |
| Muse Spark 1.2 | 27.2 |
| GPT-6 Sol | 27.1 |
| Gemini 3.7 Flash | 26.5 |
| Muse Wonder | 25.5 |
| Grok 4.5 | 25.3 |
| Muse Spark 1.3 | 25.0 |
| Muse Spark Open | 24.6 |
| Gemini 3.6 Flash | 22.1 |
| GPT-5.6 Sol | 22.0 |
| Gemini 3.5 Flash | 21.2 |
| JT-4.1 Flash 236B A21B | 20.6 |
| GPT-5.5 | 20.5 |
| Kimi K3 | 19.7 |
| Grok 4.3 | 18.0 |
| Sonnet 5 | 16.4 |
| Step 5 Preview | 16.4 |
| MiniMax M3.1 Flash | 15.6 |
| Gemini 3 Pro Preview | 15.3 |
| Grok 4.20 | 14.8 |
| GLM-5.3 | 14.3 |
| Opus 4.5 | 14.0 |
| Opus 4.6 | 13.7 |
| Qwen3.7-Max (source unspecified) | 13.5 |
| Sonnet 4.6 | 12.2 |