SAI LabsBook a demo

Safety IndexEvaluationsAA-Omniscience hallucination rate (lab-reported)

AA-Omniscience hallucination rate (lab-reported)

Interaction and honesty · Factuality and hallucinationDeveloper-reported

The developer's own AA-Omniscience hallucination rate on the full 3,000 (v1.1) or 6,000 (v1)-question set.

Metric
%. Lower is better.
Dataset
artificialanalysis.ai/evaluations/omniscience
Grader
the developer's own grader (see the model's system card)
Status
–. Measured on 2 models, 0 readings by SAI.
Notes
From Google's Gemini 4 Argon Safety & Security Report. Developer-reported; never counted.

Results

ModelValueScoreSourceSetting
Gemini 3.1 Pro50.9%n = 6,000developer-reportedGoogle, Gemini 4 Argon Safety & Security Report
Developer-reported · 2026-10
Reported by the model's developer in its own system card or report.
Full configurationfull 6,000-question set
Gemini 4 Argon15.1%n = 6,000developer-reportedGoogle, Gemini 4 Argon Safety & Security Report
Developer-reported · 2026-10
Reported by the model's developer in its own system card or report.
Full configurationfull 6,000-question set