Benchmarks · knowledge
AA-Omniscience v1.1 Non-hallucination
AA-Omniscience v1.1: one minus the hallucination rate (the share of wrong answers among questions the model did not decline). Not the combined Omniscience Index.
28 models · 76 results · data as of 2026-10-29 · artificialanalysis.ai ↗
Leaderboard
Gemini 4 Argon high
All results
76 rows| # | Model | Effort | Score | Subscription $ / task | API $ / task |
|---|---|---|---|---|---|
| 1 | Gemini 4 Argon | high | 95.7% | $0.006 | $0.018 |
| 2 | Gemini 4 Carbon | high | 88.6% | $0.25 | $0.25 |
| 3 | Qwen 3.8 27B | none | 82.0% | ||
| 4 | Grok 4.7 | xhigh | 80.0% | $0.003 | $0.028 |
| 5 | Grok 4.7 | high | 79.8% | $0.002 | $0.022 |
| 6 | Muse Spark 1.3 | xhigh | 77.5% | ||
| 7 | Muse Spark 1.3 | max | 74.8% | $0.011 | $0.023 |
| 8 | Gemini 3.5 Flash-Lite | high | 73.3% | $0.003 | $0.003 |
| 9 | Qwen 3.8 27B | xhigh | 71.4% | $0.001 | $0.010 |
| 10 | Grok 4.7 | low | 66.2% | $0.001 | $0.010 |
| 11 | Haiku 5.5 | max | 65.8% | $0.000 | $0.006 |
| 12 | Qwen 3.8 2.4T-A95B | 65.5% | |||
| 13 | GPT-6.1 Terra | max | 65.5% | $0.043 | $0.043 |
| 14 | GPT-6.1 Terra | xhigh | 63.2% | $0.008 | $0.008 |
| 15 | GPT-6.1 Terra | high | 63.0% | $0.003 | $0.003 |
| 16 | GLM 5.3 Flash | max | 61.7% | ||
| 17 | GLM 5.3 | max | 61.1% | $0.002 | $0.021 |
| 18 | Haiku 5.5 | high | 61.1% | $0.000 | $0.000 |
| 19 | Haiku 5.5 | xhigh | 60.8% | $0.000 | $0.001 |
| 20 | GPT-6.1 Terra | medium | 60.4% | $0.002 | $0.002 |
| 21 | Haiku 5.5 | medium | 59.6% | $0.000 | $0.000 |
| 22 | Haiku 5.5 | low | 59.5% | $0.000 | $0.000 |
| 23 | GPT-6.1 Terra | low | 58.4% | $0.001 | $0.001 |
| 24 | GPT-6 Astra | high | 57.5% | $0.002 | $0.028 |
| 25 | Sonnet 5.5 | max | 56.7% | $0.002 | $0.15 |
| 26 | GPT-6 Astra | low | 55.8% | $0.001 | $0.007 |
| 27 | GPT-6 Astra | xhigh | 55.4% | $0.004 | $0.052 |
| 28 | GPT-6 Astra | medium | 54.8% | $0.001 | $0.011 |
| 29 | Sonnet 5.5 | low | 52.3% | $0.000 | $0.003 |
| 30 | GPT-6.1 Sol | xhigh | 52.0% | $0.001 | $0.009 |