SAI LabsBook a demo

Safety IndexEvaluationsOver-refusal helpfulness (lab-reported)

Over-refusal helpfulness (lab-reported)

Refusal calibration · Over-refusalDeveloper-reported

The developer's own single-turn helpfulness on benign-but-sensitive prompts.

Metric
% helpful. Higher is better.
Dataset
–
Grader
the developer's own grader (see the model's system card)
Status
–. Measured on 1 model, 0 readings by SAI.
Notes
From Google's Gemini 4 Argon Safety & Security Report. Developer-reported; never counted.

Results

ModelValueScoreSourceSetting
Gemini 4 Argon96.6%developer-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 configurationHelpfulness (single-turn), internal set