Model profile
GPT 5 Pro
OpenAIdeveloper
2025-08-07release date
#36 / 267overall rank
3eval lineages
Evidence summary
GPT 5 Pro has an estimated overall rank of #36; its 90% source-sensitivity interval is #2–#232. Its behavior-only rank is #43; company governance moves the combined estimate to #36. Published evidence spans 3 evals and 4 of 7 behavior components. Its strongest relative result is SpeciEval (sea_animal_4ns, #7 of 102); its weakest is FORTRESS (over_refusal_score, #42 of 46).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- OpenRouter ↗openai/gpt-5-pro
- Official model documentation ↗Family-level model document · openai · first party
- Release source ↗curated outward evidence
Published eval results
Rank is within that sub-eval. Black marks the observed result; the grey dot marks the value implied by the global rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Better | Source |
|---|---|---|---|---|---|
| FORTRESSaverage_risk_score | #7 / 49 | 15.2 | ↓ lower | Source ↗official | |
| FORTRESSover_refusal_score | #42 / 46 | 12.43 | ↓ lower | Source ↗official | |
| MASKlying_probability_pct | #11 / 53 | 14.01 | ↓ lower | Source ↗official | |
| SpeciEvalbelief_animal_sentience | #43 / 102 | 6.83 | ↑ higher | Source ↗official | |
| SpeciEvalland_animal_4ns | #16 / 102 | 4.25 | ↓ lower | Source ↗official | |
| SpeciEvalsea_animal_4ns | #7 / 102 | 4.25 | ↓ lower | Source ↗official | |
| SpeciEvalspeciesism | #11 / 102 | 1.45 | ↓ lower | Source ↗official |