Model profile
GPT 4.1 Mini
OpenAIdeveloper
2025-04-14release date
#101 / 267overall rank
17eval lineages
Evidence summary
GPT 4.1 Mini has an estimated overall rank of #101; its 90% source-sensitivity interval is #14–#178. Its behavior-only rank is #113; company governance moves the combined estimate to #101. Published evidence spans 17 evals and 7 of 7 behavior components. Its strongest relative result is HELM Safety (simple_safety_tests, #1 of 80); its weakest is LiveSecBench (factuality, #40 of 43).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗gpt-4-1-mini
- OpenRouter ↗openai/gpt-4.1-mini
- Official model documentation ↗Family-level model document · openai · first party
- Release source ↗direct research preview identity date
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 |
|---|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #171 / 311 | 0.8198 | ↓ lower | Source ↗official | |
| AIRBench 2024 Safety Scenariossafety_scenarios | #50 / 80 | 0.604 | ↑ higher | Source ↗official | |
| BlueBench AttaQ-100attaq_harmlessness_reward_pct | #8 / 18 | 85.96 | ↑ higher | Source ↗official | |
| Cisco AI Defense Rolling Single-Turn Leaderboardsingle_turn_attack_success_rate | #93 / 105 | 72.73 | ↓ lower | Source ↗official | |
| Contextual MoralChoicehuman_agreement | #12 / 22 | 0.4 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #134 / 260 | 15.5 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #121 / 260 | 88.67 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #163 / 260 | 58.33 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #162 / 258 | 95.23 | ↑ higher | Source ↗official | |
| FORTRESSaverage_risk_score | #43 / 49 | 59.19 | ↓ lower | Source ↗official | |
| FORTRESSover_refusal_score | #6 / 46 | 1.38 | ↓ lower | Source ↗official | |
| HELM Safetyanthropic_red_team | #27 / 80 | 0.993 | ↑ higher | Source ↗official | |
| HELM Safetybbq | #47 / 80 | 0.921 | ↑ higher | Source ↗official | |
| HELM Safetyharmbench | #27 / 80 | 0.856 | ↑ higher | Source ↗official | |
| HELM Safetysimple_safety_tests | #1 / 80 | 1 | ↑ higher | Source ↗official | |
| HELM Safetyxstest | #17 / 80 | 0.974 | ↑ higher | Source ↗official | |
| LiveSecBenchethics | #35 / 43 | 23.73 | ↑ higher | Source ↗official | |
| LiveSecBenchfactuality | #40 / 43 | 22.2 | ↑ higher | Source ↗official | |
| LiveSecBenchlegality | #35 / 43 | 19.69 | ↑ higher | Source ↗official | |
| LiveSecBenchprivacy | #39 / 43 | 18.39 | ↑ higher | Source ↗official | |
| LiveSecBenchpsychological_health | #32 / 43 | 30.95 | ↑ higher | Source ↗official | |
| MASKlying_probability_pct | #40 / 53 | 50 | ↓ lower | Source ↗official | |
| PHAREbias_resistance_diagnostic | #1 / 66 | 0.8812 | ↑ higher | Source ↗official | |
| PHAREhallucination_resistance_diagnostic | #52 / 70 | 0.6781 | ↑ higher | Source ↗official | |
| PHAREharm_resistance_diagnostic | #61 / 70 | 0.8339 | ↑ higher | Source ↗official | |
| PHAREjailbreak_resistance_diagnostic | #47 / 67 | 0.4255 | ↑ higher | Source ↗official | |
| TACbase_welfare_rate | #28 / 68 | 29.49 | ↑ higher | Source ↗self run | |
| UAVBench safety-critical decision recognitionethical_safety_critical_accuracy | #17 / 27 | 0.67 | ↑ higher | Source ↗official |