Model profile
Gemini 3.1 Pro
Evidence summary
Safety. Gemini 3.1 Pro has an estimated Safety rank of #41; its 90% source-sensitivity interval is #21–#107. Its behavior-only rank is #46; company governance moves the combined estimate to #41. Published Safety evidence spans 6 eval lineages and 5 of 7 components. Its strongest relative result is PHARE (hallucination_resistance_diagnostic, #3 of 70); its weakest is Every Model Cheats — Cybench Cheat Propensity (severe_anti_cheat_cheat_propensity_pct, #17 of 22).
Freedom. Gemini 3.1 Pro has an estimated Freedom rank of #422; its 90% source-sensitivity interval is #272–#506. Published Freedom evidence spans 1 eval lineages and 1 of 1 components. Its strongest relative result is PHARE (jailbreak_resistance_diagnostic, #37 of 67); its weakest is PHARE (harm_resistance_diagnostic, #45 of 70).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Official model documentation ↗Family-level model document · google · first party
- Release source ↗direct research preview identity date
Safety evals
Rank and direction are specific to the Safety portfolio. Black marks the observed result; the grey dot marks the value implied by the Safety rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| Arena Factuality — Search Arena (factuality-only weighting)factuality_bt_rating | #18 / 30 | ↑1182.0 | Source ↗official | |
| DystopiaBenchbasaglia_score | #18 / 50 | ↓55.37 | Source ↗official | |
| DystopiaBenchbaudrillard_score | #26 / 50 | ↓57.8 | Source ↗official | |
| DystopiaBenchhuxley_score | #14 / 50 | ↓49.97 | Source ↗official | |
| DystopiaBenchlaguardia_score | #19 / 50 | ↓61.97 | Source ↗official | |
| DystopiaBenchorwell_score | #21 / 50 | ↓59.17 | Source ↗official | |
| DystopiaBenchpetrov_score | #16 / 50 | ↓61.67 | Source ↗official | |
| Every Model Cheats — Cybench Cheat Propensitybaseline_cheat_propensity_pct | #2 / 22 | ↓13 | Source ↗official | |
| Every Model Cheats — Cybench Cheat Propensitysevere_anti_cheat_cheat_propensity_pct | #17 / 22 | ↓13 | Source ↗official | |
| Every Model Cheats — Cybench Cheat Propensitystandard_anti_cheat_cheat_propensity_pct | #10 / 22 | ↓13 | Source ↗official | |
| HarmVideoBenchharmful_video_safety_recognition_reasoning | #1 / 19 | ↑0.821 | Source ↗official | |
| PHAREbias_resistance_diagnostic | #27 / 66 | ↑0.481 | Source ↗official | |
| PHAREhallucination_resistance_diagnostic | #3 / 70 | ↑0.8759 | Source ↗official | |
| PHAREharm_resistance_diagnostic | #26 / 70 | ↑0.9519 | Source ↗official | |
| PHAREjailbreak_resistance_diagnostic | #31 / 67 | ↑0.5164 | Source ↗official | |
| The Dictatorship Evaloverall_resistance_rate | #12 / 20 | ↑53.4 | Source ↗official |
Freedom evals
Rank and direction are specific to the Freedom portfolio. Black marks the observed result; the grey dot marks the value implied by the Freedom rank. Values stay on each source’s native scale.