Model profile
Grok 4 Fast
Evidence summary
Grok 4 Fast has an estimated overall rank of #102; its 90% source-sensitivity interval is #40–#174. Its behavior-only rank is #100; company governance moves the combined estimate to #102. Published evidence spans 12 evals and 5 of 7 behavior components. Its strongest relative result is PHARE (bias_resistance_diagnostic, #2 of 66); its weakest is CAIS Risk Index (mask, #49 of 51).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗grok-4-fast
- OpenRouter ↗x-ai/grok-4-fast
- 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 | #83 / 311 | 0.6596 | ↓ lower | Source ↗official | |
| ANIMAscore | #13 / 19 | 0.6164 | ↑ higher | Source ↗official | |
| BrokenMathsycophancy | #4 / 9 | 40 | ↓ lower | Source ↗official | |
| CAIS Risk Indexbioweapons_assistance | #19 / 48 | 64.1 | ↓ lower | Source ↗official | |
| CAIS Risk Indexhle_overconfidence | #35 / 49 | 68.7 | ↓ lower | Source ↗official | |
| CAIS Risk Indexmachiavelli | #41 / 45 | 96.2 | ↓ lower | Source ↗official | |
| CAIS Risk Indexmask | #49 / 51 | 70 | ↓ lower | Source ↗official | |
| CAIS Risk Indexpolitical_manipulation | #2 / 32 | 30.8 | ↓ lower | Source ↗official | |
| CAIS Risk Indextextquests_harm | #29 / 48 | 19.8 | ↓ lower | Source ↗official | |
| Cisco AI Defense Rolling Single-Turn Leaderboardsingle_turn_attack_success_rate | #43 / 105 | 30.03 | ↓ lower | Source ↗official | |
| MACHIAVELLIdeception_relative_random_pct | #43 / 50 | 96.2 | ↓ lower | Source ↗official | |
| PHAREbias_resistance_diagnostic | #2 / 66 | 0.8026 | ↑ higher | Source ↗official | |
| PHAREhallucination_resistance_diagnostic | #62 / 70 | 0.634 | ↑ higher | Source ↗official | |
| PHAREharm_resistance_diagnostic | #64 / 70 | 0.8134 | ↑ higher | Source ↗official | |
| PHAREjailbreak_resistance_diagnostic | #49 / 67 | 0.412 | ↑ higher | Source ↗official | |
| UAVBench safety-critical decision recognitionethical_safety_critical_accuracy | #14 / 27 | 0.695 | ↑ 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.