Versions

One version is on record: the copy we retrieved on 26 Sep 2026. We know of no other.

  1. 26 Sep 2026

    Date retrieved

    Copy retrieved 26 Sep 2026

    Date from
    the date we retrieved it; the copy states no version date
    Copy
    Copy retrieved on 26 Sep 2026
    Values
    17 values recorded from this version

No version has a file hash or an archived snapshot yet. From dataset v0.2 each retrieved version carries both (Methodology §7).

Revisions

No revisions are recorded for this document. We know of only one version of it.

Values

Every value we recorded from this document, grouped by metric family and ordered by where the document prints it. Location is the section, table or page as the document numbers it. 1 of the 17 has been blind-verified: a second reader found the same value without seeing ours.

KeyVerifiedUnverifiedDisputedCorrected read off a figure or stated in wordsA value opens its source and history.

M1 Evaluation awareness

1 value

M1 Evaluation awareness: values in the Gemini 3.1 Pro Model Card
ModelEvaluationConditionValueLocationChecked
Gemini 3.1 ProSituational awareness challengesSuccess rate on three hardest challengesDeep ThinkFrontier Safety: misalignmentUnverified

M3 Sabotage and sandbagging

1 value

M3 Sabotage and sandbagging: values in the Gemini 3.1 Pro Model Card
ModelEvaluationConditionValueLocationChecked
Gemini 3.1 ProStealth challengesComparison with Gemini 3 ProDeep ThinkFrontier Safety: misalignmentUnverified

M4 Misalignment audits

1 value

M4 Misalignment audits: values in the Gemini 3.1 Pro Model Card
ModelEvaluationConditionValueLocationChecked
Gemini 3.1 ProFrontier Safety Framework determination (misalignment)Instrumental reasoningDeep ThinkFrontier Safety tableUnverified

M7 Harmful compliance and over-refusal

6 values

M7 Harmful compliance and over-refusal: values in the Gemini 3.1 Pro Model Card
ModelEvaluationConditionValueLocationChecked
Gemini 3.1 ProChild safety launch thresholdsLaunch thresholds metNone statedSafety evaluationsUnverified
Gemini 3.1 ProImage to Text SafetyImage-to-text policy-violation rate deltavs Gemini 3 ProSafety evaluations tableUnverified
Gemini 3.1 ProMultilingual SafetyMultilingual policy-violation rate deltavs Gemini 3 ProSafety evaluations tableUnverified
Gemini 3.1 ProText to Text SafetyPolicy-violation rate deltavs Gemini 3 ProSafety evaluations tableVerified
Gemini 3.1 ProToneObjective tone of refusals deltavs Gemini 3 ProSafety evaluations tableUnverified
Gemini 3.1 ProUnjustified refusalsUnjustified refusals deltavs Gemini 3 ProSafety evaluations tableUnverified

M10 Dangerous capabilities and risk determinations

8 values

M10 Dangerous capabilities and risk determinations: values in the Gemini 3.1 Pro Model Card
ModelEvaluationConditionValueLocationChecked
Gemini 3.1 ProFrontier Safety Framework determinationCBRN Uplift Level 1Deep ThinkFrontier Safety tableUnverified
Gemini 3.1 ProFrontier Safety Framework determinationCyber Uplift Level 1None statedFrontier Safety tableUnverified
Gemini 3.1 ProFrontier Safety Framework determinationHarmful manipulationDeep ThinkFrontier Safety tableUnverified
Gemini 3.1 ProFrontier Safety Framework determinationML R&DDeep ThinkFrontier Safety tableUnverified
Gemini 3 ProHarmful manipulation efficacy studyMax odds ratio vs non-AI baselineNone statedFrontier Safety: manipulationUnverified
Gemini 3.1 ProHarmful manipulation efficacy studyMax odds ratio vs non-AI baselineDeep ThinkFrontier Safety: manipulationUnverified
Gemini 3 ProRE-BenchAverage normalised scoreNone statedFrontier Safety: ML R&DUnverified
Gemini 3.1 ProRE-BenchAverage normalised scoreDeep ThinkFrontier Safety: ML R&DUnverified

Extraction coverage

What we read of this document, and where each value was read.

Our note We have not written a coverage note for this document yet.

All 17 values were read in the document itself.

Of the 17 values, 2 are statements in words rather than numbers; they are marked * and left out of charts by default.

All 17 values were extracted for version 0 of the dataset through a web reader, which did not always reach the later sections of long PDFs (Methodology §3).