Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Knowledge benchmark
Measuring Short-Form Factuality in Large Language Models. Every model the catalog carries a published SimpleQA value for, ranked by that value.
A benchmark that evaluates the ability of language models to answer short, fact-seeking questions accurately. Focuses on factual correctness rather than reasoning complexity. 18 models in the ModelChoice catalog have a published value.
Each point is a model with a published SimpleQA value and complete token prices, on the 75% input / 25% output blend. Frontier models are spaced evenly along the cost axis.
18 models with a published SimpleQA value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Short-form Q&A |
|---|---|---|---|
| 1 | OpenAI | 62.5 | |
| 2 | DeepSeek | 57.9 | |
| 3 | Alibaba | 54.3 |
What the benchmark measures and how its value is published, from the source’s own record.
SimpleQA (Measuring Short-Form Factuality in Large Language Models) is a knowledge benchmark published by Measuring short-form factuality in large language models. A benchmark that evaluates the ability of language models to answer short, fact-seeking questions accurately. Focuses on factual correctness rather than reasoning complexity. Its evaluation set is described by the source as factual questions. The source rates its difficulty as factual accuracy focused.
Each model's result is published by Measuring short-form factuality in large language models as short-form Q&A. ModelChoice prints that value exactly as published, and never converts or rescales it. This page ranks the 18 models that carry a published value, highest value first; a model the source never scored is left out rather than ranked at zero. The source does not state whether a higher value is the better result, so neither does this page.
| 4 | xAI | 43.6 |
| 5 | DeepSeek | 34.1 |
| 6 | Moonshot AI | 31 |
| 6 | Microsoft | 31 |
| 8 | DeepSeek | 30.1 |
| 9 | 29.9 |
| 10 | DeepSeek | 27.8 |
| 11 | DeepSeek | 24.9 |
| 12 | MiniMax | 23.7 |
| 13 | OpenAI | 20.2 |
| 14 | 16.5 |
| 15 | 10.7 |
| 16 | Mistral | 10.43 |
| 17 | 8.6 |
| 18 | OpenAI | 7.6 |
Evidence key: ObservedLast good
Rows are ordered by the value Measuring short-form factuality in large language models published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.