AA-Omniscience Index: Results as Published

AA-Omniscience measures factual recall and whether a model admits a gap instead of guessing. It contains 6,000 questions on 42 topics in six domains, drawn from academic and industry sources. The Omniscience Index runs from -100 to 100: a correct answer raises it, a hallucination lowers it, a refusal is not penalized, and 0 means the model is right as often as it is wrong. The page also reports accuracy, the share of all questions answered correctly, and a hallucination rate among responses that were not correct.

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.