Model profile
Gemini 2.5 Pro Exp
Evidence summary
Safety. Gemini 2.5 Pro Exp has an estimated Safety rank of #38; its 90% source-sensitivity interval is #4–#195. Its behavior-only rank is #39; company governance moves the combined estimate to #38. Published Safety evidence spans 4 eval lineages and 3 of 7 components. Its strongest relative result is Confabulations (confabulation_rate, #3 of 52); its weakest is FORTRESS (average_risk_score, #40 of 50).
Freedom. Gemini 2.5 Pro Exp has an estimated Freedom rank of #140; its 90% source-sensitivity interval is #86–#430. Published Freedom evidence spans 1 eval lineages and 1 of 1 components. Its strongest relative result is FORTRESS (over_refusal_score, #6 of 49); its weakest is FORTRESS (average_risk_score, #11 of 50).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗gemini-2-5-pro-03-25
- OpenRouter ↗google/gemini-2.5-pro-exp-03-25
- 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 |
|---|---|---|---|---|
| CAIS Risk Indexhle_overconfidence | #42 / 55 | ↓71 | Source ↗official | |
| CAIS Risk Indexmask | #44 / 57 | ↓46.9 | Source ↗official | |
| Confabulationsconfabulation_rate | #3 / 52 | ↓3.96 | Source ↗official | |
| FORTRESSaverage_risk_score | #40 / 50 | ↓54.89 | Source ↗official | |
| FORTRESSover_refusal_score | #6 / 49 | ↓1.38 | Source ↗official | |
| Humanity's Last Exam RMS calibration error (Scale Labs)calibrationError | #22 / 42 | ↓71 | Source ↗official | |
| MASKlying_probability_pct | #33 / 53 | ↓44.07 | 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.