Model profile
o1 Preview
Evidence summary
Safety. o1 Preview has an estimated Safety rank of #52; its 90% source-sensitivity interval is #4–#197. Its behavior-only rank is #51; company governance moves the combined estimate to #52. Published Safety evidence spans 4 eval lineages and 6 of 7 components. Its strongest relative result is Enkrypt AI Safety Leaderboard (cbrn_attack_non_success_rate, #14 of 272); its weakest is SORRY-Bench (avg, #22 of 51).
Freedom. o1 Preview has an estimated Freedom rank of #615; its 90% source-sensitivity interval is #384–#656. Published Freedom evidence spans 3 eval lineages and 1 of 1 components. Its strongest relative result is SORRY-Bench (avg, #30 of 51); its weakest is Enkrypt AI Safety Leaderboard (cbrn_attack_non_success_rate, #258 of 272).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗o1-preview
- OpenRouter ↗openai/o1-preview
- System card ↗Family-level model document · OpenAI · 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 |
|---|---|---|---|---|
| Confabulationsconfabulation_rate | #21 / 52 | ↓18.32 | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #80 / 272 | ↑23.51 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #14 / 272 | ↑95.5 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #26 / 272 | ↑98.33 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #91 / 270 | ↑97.64 | Source ↗official | |
| Large-scale Moral Machine experiment on LLMshuman_choice_distance | #11 / 39 | ↓0.8621 | Source ↗official | |
| SORRY-Benchavg | #22 / 51 | ↓0.25 | 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.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #258 / 272 | ↓95.5 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #241 / 272 | ↓98.33 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #177 / 270 | ↓97.64 | Source ↗official | |
| SORRY-Benchavg | #30 / 51 | ↑0.25 | Source ↗official | |
| SpeechMap model completioncomplete_pct | #153 / 181 | ↑35.8 | Source ↗official |