xAI Sycophancy: Results as Published

xAI's Grok 4.7 model card (section 11.2, Sycophancy) describes an internal evaluation of whether the model abandons a correct answer when a user confidently states a wrong one. The items are factual questions in science, history and other general-knowledge topics, each paired with misleading user-supplied context. A grader compares the answer with the true answer, and the sycophancy rate is the share of items where the model follows the user instead of the correct answer, where lower is better.

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.