Model profile
Grok 3 Mini
xAIdeveloper
2025-04-03release date
#87 / 267overall rank
15eval lineages
Evidence summary
Grok 3 Mini has an estimated overall rank of #87; its 90% source-sensitivity interval is #28–#184. Its behavior-only rank is #91; company governance moves the combined estimate to #87. Published evidence spans 15 evals and 7 of 7 behavior components. Its strongest relative result is AA-Omniscience (hallucination_rate, #10 of 311); its weakest is Cisco AI Defense Rolling Single-Turn Leaderboard (single_turn_attack_success_rate, #105 of 105).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗grok-3-mini-reasoning
- OpenRouter ↗x-ai/grok-3-mini
- Official model documentation ↗Family-level model document · xai · 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 | #10 / 311 | 0.2535 | ↓ lower | Source ↗official | |
| AIRBench 2024 Safety Scenariossafety_scenarios | #61 / 80 | 0.535 | ↑ higher | Source ↗official | |
| Cisco AI Defense Rolling Single-Turn Leaderboardsingle_turn_attack_success_rate | #105 / 105 | 86.64 | ↓ lower | Source ↗official | |
| Confabulationsconfabulation_rate | #9 / 52 | 8.911 | ↓ lower | Source ↗official | |
| Emergent Collusionhigh_illegality_game_rate | #4 / 13 | 0.3 | ↓ lower | Source ↗official | |
| HELM Safetyanthropic_red_team | #18 / 80 | 0.995 | ↑ higher | Source ↗official | |
| HELM Safetybbq | #13 / 80 | 0.967 | ↑ higher | Source ↗official | |
| HELM Safetyharmbench | #64 / 80 | 0.572 | ↑ higher | Source ↗official | |
| HELM Safetysimple_safety_tests | #29 / 80 | 0.993 | ↑ higher | Source ↗official | |
| HELM Safetyxstest | #6 / 80 | 0.984 | ↑ higher | Source ↗official | |
| LiveSecBenchethics | #39 / 43 | 20.84 | ↑ higher | Source ↗official | |
| LiveSecBenchfactuality | #29 / 43 | 37.05 | ↑ higher | Source ↗official | |
| LiveSecBenchlegality | #33 / 43 | 20.79 | ↑ higher | Source ↗official | |
| LiveSecBenchprivacy | #32 / 43 | 30.02 | ↑ higher | Source ↗official | |
| LiveSecBenchpsychological_health | #31 / 43 | 33.15 | ↑ higher | Source ↗official | |
| PacifAIstp_score | #6 / 7 | 79.77 | ↑ higher | Source ↗official | |
| PHAREbias_resistance_diagnostic | #32 / 66 | 0.464 | ↑ higher | Source ↗official | |
| PHAREhallucination_resistance_diagnostic | #48 / 70 | 0.6862 | ↑ higher | Source ↗official | |
| PHAREharm_resistance_diagnostic | #49 / 70 | 0.9047 | ↑ higher | Source ↗official | |
| PHAREjailbreak_resistance_diagnostic | #60 / 67 | 0.3436 | ↑ higher | Source ↗official | |
| SOSBenchbiology_pvr | #16 / 23 | 0.758 | ↓ lower | Source ↗official | |
| SOSBenchchemistry_pvr | #16 / 23 | 0.586 | ↓ lower | Source ↗official | |
| SOSBenchmedicine_pvr | #16 / 23 | 0.746 | ↓ lower | Source ↗official | |
| SOSBenchpharmacology_pvr | #17 / 23 | 0.93 | ↓ lower | Source ↗official | |
| SOSBenchphysics_pvr | #16 / 23 | 0.708 | ↓ lower | Source ↗official | |
| SOSBenchpsychology_pvr | #15 / 23 | 0.7 | ↓ lower | Source ↗official | |
| SpeciEvalbelief_animal_sentience | #49 / 102 | 6.815 | ↑ higher | Source ↗official | |
| SpeciEvalland_animal_4ns | #50 / 102 | 4.545 | ↓ lower | Source ↗official | |
| SpeciEvalsea_animal_4ns | #71 / 102 | 4.91 | ↓ lower | Source ↗official | |
| SpeciEvalspeciesism | #24 / 102 | 1.68 | ↓ lower | Source ↗official |