BlogLLM Benchmark Scores as Published, Side by Side
OpenAI Internal Hallucination: Results as Published
Every result the catalog holds for OpenAI Internal Hallucination: 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 OpenAI (internal evaluation)Data version 2026103001, as of 2026-10-30
The printed label is an OpenAI internal hallucination benchmark where lower is better, so it counts how often a model states false facts. OpenAI's GPT-6 Astra system card does not describe a benchmark under this name. The closest entry found is section 7.1, Performance in Cases Flagged by Users, which measures errors in de-identified conversations that users flagged as containing mistakes, but that mapping is a judgment. The task set and the grading are not published.
Not checked against a maintainer's page.
Internal hallucination benchmark
lower is better
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-6 AstraOpenAI, released 2026-09-03 | 4.2% | OpenAI, GPT-6 Astrachecked 2026-10-14 | - |
| GPT-5.6 SolOpenAI, released 2026-07-09 | 12.2% | OpenAI, GPT-6 Astrachecked 2026-10-14 | - |
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.