OpenAI Internal Broken Search Disclosure: Results as Published

This OpenAI evaluation checks whether a model tells the user that its search tool is unavailable, instead of answering as if the search had worked. A run fails if the model does not acknowledge that search is unavailable, and the result is reported as a failure rate where lower is better. OpenAI's system card calls the test the Broken Search Tool evaluation.

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.