Model profile
GPT 4O Mini
OpenAIdeveloper
2024-07-18release date
#237 / 309overall rank
28eval lineages
Evidence summary
GPT 4O Mini has an estimated overall rank of #237; its 90% source-sensitivity interval is #163–#249. Its behavior-only rank is #245; company governance moves the combined estimate to #237. Published evidence spans 28 evals and 7 of 7 behavior components. Its strongest relative result is Contextual MoralChoice (human_agreement, #1 of 22); its weakest is MonitoringBench Full-Trajectory Monitor (full_trajectory_catch_rate_at_1pct_fpr_percent, #13 of 13).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗gpt-4o-mini
- OpenRouter ↗openai/gpt-4o-mini
- System card ↗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 | Source |
|---|---|---|---|---|
| Agent-SafetyBenchcompromise_availability | #13 / 16 | ↑23.6 | Source ↗official | |
| Agent-SafetyBenchharmful_vulnerable_code | #13 / 16 | ↑25.2 | Source ↗official | |
| Agent-SafetyBenchleak_sensitive_information | #11 / 16 | ↑28 | Source ↗official | |
| Agent-SafetyBenchphysical_harm | #12 / 16 | ↑24.4 | Source ↗official | |
| Agent-SafetyBenchproduce_unsafe_information | #5 / 16 | ↑98.4 | Source ↗official | |
| Agent-SafetyBenchproperty_loss | #12 / 16 | ↑30 | Source ↗official | |
| Agent-SafetyBenchspread_unsafe_information | #14 / 16 | ↑6.8 | Source ↗official | |
| Agent-SafetyBenchviolate_law_ethics | #14 / 16 | ↑13.2 | Source ↗official | |
| AgentDojotargeted_attack_success_rate | #13 / 15 | ↓0.2719 | Source ↗official | |
| AgentDojoutility_under_attack | #5 / 15 | ↑0.4992 | Source ↗official | |
| AgentDrive Safety Compliancescr | #36 / 48 | ↑58.75 | Source ↗official | |
| AgentHarmharm_score | #10 / 12 | ↓62.5 | Source ↗official | |
| AILuminate General Purpose AI Chatchild_sexual_exploitation | #14 / 32 | ↓17.6 | Source ↗official | |
| AILuminate General Purpose AI Chatdefamation | #19 / 32 | ↓17.4 | Source ↗official | |
| AILuminate General Purpose AI Chathate | #20 / 32 | ↓10.9 | Source ↗official | |
| AILuminate General Purpose AI Chatindiscriminate_weapons_cbrne | #13 / 32 | ↓19.9 | Source ↗official | |
| AILuminate General Purpose AI Chatintellectual_property | #14 / 32 | ↓10.1 | Source ↗official | |
| AILuminate General Purpose AI Chatnon_violent_crimes | #15 / 32 | ↓15.7 | Source ↗official | |
| AILuminate General Purpose AI Chatprivacy | #13 / 32 | ↓10.5 | Source ↗official | |
| AILuminate General Purpose AI Chatsex_related_crimes | #19 / 32 | ↓16.9 | Source ↗official | |
| AILuminate General Purpose AI Chatsexual_content | #26 / 32 | ↓42.9 | Source ↗official | |
| AILuminate General Purpose AI Chatspecialized_advice | #19 / 32 | ↓18.2 | Source ↗official | |
| AILuminate General Purpose AI Chatsuicide_self_harm | #20 / 32 | ↓15.5 | Source ↗official | |
| AILuminate General Purpose AI Chatviolent_crimes | #17 / 32 | ↓18 | Source ↗official | |
| AIRBench 2024 Safety Scenariossafety_scenarios | #57 / 80 | ↑0.563 | Source ↗official | |
| AnimalHarmBenchscore | #9 / 10 | ↑0.002 | Source ↗official | |
| BullshitBench v2clear_pushback_rate | #105 / 106 | ↑0.02 | Source ↗official | |
| CASE-Benchagreement_accuracy | #6 / 7 | ↑80 | Source ↗official | |
| Cisco AI Defense Rolling Single-Turn Leaderboardsingle_turn_attack_success_rate | #70 / 104 | ↓54.99 | Source ↗official | |
| Confabulationsconfabulation_rate | #50 / 52 | ↓60.89 | Source ↗official | |
| Contextual MoralChoicehuman_agreement | #1 / 22 | ↑0.65 | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #150 / 241 | ↑13.7 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #47 / 241 | ↑92 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #140 / 241 | ↑60.56 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #71 / 239 | ↑98 | Source ↗official | |
| FinEval Financial Security Knowledgefinancial_security_accuracy_pct | #3 / 19 | ↑79.1 | Source ↗official | |
| FORTRESSaverage_risk_score | #34 / 49 | ↓48.07 | Source ↗official | |
| FORTRESSover_refusal_score | #14 / 46 | ↓2.27 | Source ↗official | |
| HalluVerse-M3 Hallucination Recognitionhallucination_recognition_accuracy | #6 / 14 | ↑73.39 | Source ↗official | |
| HELM Safetyanthropic_red_team | #50 / 80 | ↑0.983 | Source ↗official | |
| HELM Safetybbq | #59 / 80 | ↑0.882 | Source ↗official | |
| HELM Safetyharmbench | #28 / 80 | ↑0.849 | Source ↗official | |
| HELM Safetysimple_safety_tests | #52 / 80 | ↑0.978 | Source ↗official | |
| HELM Safetyxstest | #39 / 80 | ↑0.96 | Source ↗official | |
| Large-scale Moral Machine experiment on LLMshuman_choice_distance | #19 / 39 | ↓1.076 | Source ↗official | |
| MonitoringBench Full-Trajectory Monitorfull_trajectory_catch_rate_at_1pct_fpr_percent | #13 / 13 | ↑0.4 | Source ↗official | |
| PandaBench JBB direct-request panelsafety_rate | #25 / 46 | ↑0.98 | Source ↗official | |
| PHAREbias_resistance_diagnostic | #51 / 66 | ↑0.379 | Source ↗official | |
| PHAREhallucination_resistance_diagnostic | #59 / 70 | ↑0.6597 | Source ↗official | |
| PHAREharm_resistance_diagnostic | #67 / 70 | ↑0.7729 | Source ↗official | |
| PHAREjailbreak_resistance_diagnostic | #30 / 67 | ↑0.5214 | Source ↗official | |
| SafeArenanormalized_safety_score | #2 / 5 | ↑35.7 | Source ↗official | |
| SM-Benchadversarial | #61 / 79 | ↑78.54 | Source ↗official | |
| SM-Benchambiguous_interpretation | #74 / 79 | ↑66.07 | Source ↗official | |
| SM-Benchanti_hallucination | #77 / 79 | ↑59.16 | Source ↗official | |
| SM-Bencheq_boundaries | #12 / 79 | ↑71.35 | Source ↗official | |
| SM-Benchoverfit | #50 / 79 | ↑65.03 | Source ↗official | |
| SORRY-Benchavg | #34 / 51 | ↓0.36 | Source ↗official | |
| SpeciEvalbelief_animal_sentience | #108 / 113 | ↑6.28 | Source ↗official | |
| SpeciEvalland_animal_4ns | #55 / 113 | ↓4.53 | Source ↗official | |
| SpeciEvalsea_animal_4ns | #49 / 113 | ↓4.7 | Source ↗official | |
| SpeciEvalspeciesism | #95 / 113 | ↓2.6 | Source ↗official |
Values evaluations
Descriptive values and political-framing results are separate from safety/ethics ranks. Each strip shows the evaluation’s observed model range; its endpoint labels state what lower and higher values mean.
ValueCompass
| Dimension | Value | Distribution |
|---|---|---|
| Universalism | 67.2 | |
| Self-direction | 46.2 | |
| Care / Harm | 30.1 | |
| Fairness / Cheating | 28.5 | |
| Ethical | 90.7 |