Enterprise AI in Numbers · Capability and cost
Top score on SimpleQA Verified, the benchmark of factual accuracy without search
76.0%
76.0% (Gemini 4 Argon) as of September 30, 2026, according to Epoch AI, Benchmarking hub, up from 75.6% on September 3, 2026. The highest share of SimpleQA Verified questions (short questions with one verifiable answer) that any evaluated model answers correctly without web search, with its lab.
What it means for leaders
Even the best model gets a large share of plain factual questions wrong from memory. Ground customer-facing and decision-support answers in your own sources rather than relying on what the model knows.
What would change the reading: Whether the best score keeps rising, and how close smaller, cheaper models get.
The trend
Every reading
| Date | Value | Change | Detail | Source |
|---|---|---|---|---|
| 76.0% | +0.4 points | Gemini 4 Argon | Source | |
| 75.6% | +2.1 points | GPT-6 Astra | Source | |
| 73.5% | +6.7 points | Gemini 3.1 Pro | Source | |
| 66.8% | +16.7 points | Gemini 3 Flash | Source | |
| 50.1% | +0.7 points | GPT-5 | Source | |
| 49.4% | +8.3 points | o3 | Source | |
| 41.1% | +15.1 points | o1 | Source | |
| 26.0% | +13.4 points | GPT-4o (Aug 2024) | Source | |
| 12.6% | — | Claude 3 Opus | Source |
Definition and method
- What it measures
- The highest share of SimpleQA Verified questions (short questions with one verifiable answer) that any evaluated model answers correctly without web search, with its lab.
- Population
- AI models evaluated on SimpleQA Verified, as compiled by Epoch AI
- Updated
- When new models are evaluated
- Source
- Epoch AI, Benchmarking hub · CC BY 4.0
Caveats
- Scores come from independent evaluators compiled by Epoch AI, not from the labs, and depend on the settings and tools each run used.
- The leader is the best result so far, dated by the release of the model that set it; a model evaluated late can take the lead after its release date.
- It measures recall without search; with retrieval from trusted sources, accuracy is higher.
Cite this
Enterprise AI in Numbers, "Top score on SimpleQA Verified, the benchmark of factual accuracy without search", https://lucianoayres.com/ai-brief/data/factual-accuracy-benchmark-leader/, retrieved October 25, 2026. Data: Epoch AI, Benchmarking hub.