Model profile
Llama 4 Scout
Evidence summary
Llama 4 Scout has an estimated overall rank of #216; its 90% source-sensitivity interval is #164–#223. Its behavior-only rank is #215; company governance moves the combined estimate to #216. Published evidence spans 16 evals and 7 of 7 behavior components. Its strongest relative result is PHARE (bias_resistance_diagnostic, #7 of 66); its weakest is Vigil Mental Health Safety (overall_score, #23 of 23).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗llama-4-scout
- OpenRouter ↗meta-llama/llama-4-scout
- Official model page ↗Family-level model document · meta · 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 | #135 / 311 | 0.7831 | ↓ lower | Source ↗official | |
| AIRBench 2024 Safety Scenariossafety_scenarios | #63 / 80 | 0.523 | ↑ higher | Source ↗official | |
| BullshitBench v2clear_pushback_rate | #67 / 105 | 0.19 | ↑ higher | Source ↗official | |
| Cisco AI Defense Rolling Single-Turn Leaderboardsingle_turn_attack_success_rate | #75 / 105 | 58.72 | ↓ lower | Source ↗official | |
| DystopiaBenchbasaglia_score | #34 / 50 | 67.87 | ↓ lower | Source ↗official | |
| DystopiaBenchbaudrillard_score | #45 / 50 | 74.3 | ↓ lower | Source ↗official | |
| DystopiaBenchhuxley_score | #24 / 50 | 70.13 | ↓ lower | Source ↗official | |
| DystopiaBenchlaguardia_score | #29 / 50 | 67.9 | ↓ lower | Source ↗official | |
| DystopiaBenchorwell_score | #23 / 50 | 68.6 | ↓ lower | Source ↗official | |
| DystopiaBenchpetrov_score | #20 / 50 | 70.47 | ↓ lower | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #160 / 260 | 13.95 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #140 / 260 | 87.83 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #138 / 260 | 68.33 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #78 / 258 | 98.05 | ↑ higher | Source ↗official | |
| HELM Safetyanthropic_red_team | #67 / 80 | 0.965 | ↑ higher | Source ↗official | |
| HELM Safetybbq | #60 / 80 | 0.875 | ↑ higher | Source ↗official | |
| HELM Safetyharmbench | #62 / 80 | 0.6 | ↑ higher | Source ↗official | |
| HELM Safetysimple_safety_tests | #58 / 80 | 0.97 | ↑ higher | Source ↗official | |
| HELM Safetyxstest | #43 / 80 | 0.956 | ↑ higher | Source ↗official | |
| PHAREbias_resistance_diagnostic | #7 / 66 | 0.671 | ↑ higher | Source ↗official | |
| PHAREhallucination_resistance_diagnostic | #69 / 70 | 0.582 | ↑ higher | Source ↗official | |
| PHAREharm_resistance_diagnostic | #65 / 70 | 0.8104 | ↑ higher | Source ↗official | |
| PHAREjailbreak_resistance_diagnostic | #35 / 67 | 0.4893 | ↑ higher | Source ↗official | |
| SOSBenchbiology_pvr | #13 / 23 | 0.488 | ↓ lower | Source ↗official | |
| SOSBenchchemistry_pvr | #12 / 23 | 0.436 | ↓ lower | Source ↗official | |
| SOSBenchmedicine_pvr | #13 / 23 | 0.688 | ↓ lower | Source ↗official | |
| SOSBenchpharmacology_pvr | #14 / 23 | 0.874 | ↓ lower | Source ↗official | |
| SOSBenchphysics_pvr | #13 / 23 | 0.492 | ↓ lower | Source ↗official | |
| SOSBenchpsychology_pvr | #13 / 23 | 0.51 | ↓ lower | Source ↗official | |
| SpeciEvalbelief_animal_sentience | #19 / 102 | 6.97 | ↑ higher | Source ↗official | |
| SpeciEvalland_animal_4ns | #92 / 102 | 5 | ↓ lower | Source ↗official | |
| SpeciEvalsea_animal_4ns | #97 / 102 | 5.33 | ↓ lower | Source ↗official | |
| SpeciEvalspeciesism | #52 / 102 | 2.02 | ↓ lower | Source ↗official | |
| UAVBench safety-critical decision recognitionethical_safety_critical_accuracy | #19 / 27 | 0.635 | ↑ higher | Source ↗official | |
| Vigil Mental Health Safetyoverall_score | #23 / 23 | 20 | ↑ higher | 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.