BlogLLM Benchmark Scores as Published, Side by Side
FActScore: Results as Published
Every result the catalog holds for FActScore: 2 models from 1 developer, read from 1 page. Each number is as printed by the company named beside it; nothing is ranked or rescored. Models are listed by developer, newest first.
Area General Knowledge and ReasoningMaintainer Sewon Min and co-authorsIntroduced 2023Data version 2026103001, as of 2026-10-30
FActScore measures how much of a long generated text is supported by a knowledge source. The generation is split into atomic facts, and each fact is checked against that source. The score is the percentage of atomic facts the source supports. The paper also describes an automatic estimator that uses retrieval and a language model in place of a human check.
FActScore (hallucination rate)
lower is better
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-5OpenAI, released 2025-08-07 | 1% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| o3OpenAI, released 2025-04-16 | 5.7% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.