Model profile
Gemma 4 E4B
Evidence summary
Safety. Gemma 4 E4B has an estimated Safety rank of #161; its 90% source-sensitivity interval is #1–#299. Its behavior-only rank is #176; company governance moves the combined estimate to #161. Published Safety evidence spans 3 eval lineages and 4 of 7 components. Its strongest relative result is AA-Omniscience (hallucination_rate, #33 of 338); its weakest is Enkrypt AI Safety Leaderboard (cbrn_attack_non_success_rate, #230 of 239).
Freedom. Gemma 4 E4B has an estimated Freedom rank of #325; its 90% source-sensitivity interval is #223–#418. Published Freedom evidence spans 2 eval lineages and 1 of 1 components. Its strongest relative result is Enkrypt AI Safety Leaderboard (cbrn_attack_non_success_rate, #9 of 239); its weakest is Enkrypt AI Safety Leaderboard (harmful_attack_non_success_rate, #172 of 239).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗gemma-4-e4b
- Official model documentation ↗Family-level model document · google · first party
- Release source ↗curated outward evidence
Safety evals
Rank and direction are specific to the Safety portfolio. Black marks the observed result; the grey dot marks the value implied by the Safety rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #33 / 338 | ↓0.3094 | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #226 / 239 | ↑6.2 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #230 / 239 | ↑66.83 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #64 / 239 | ↑85.56 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #71 / 237 | ↑98 | Source ↗official | |
| MT-JailBench CrescendoXsafety_score | #11 / 21 | ↑21.38 | Source ↗official |
Freedom evals
Rank and direction are specific to the Freedom portfolio. Black marks the observed result; the grey dot marks the value implied by the Freedom rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #9 / 239 | ↓66.83 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #172 / 239 | ↓85.56 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #165 / 237 | ↓98 | Source ↗official | |
| MT-JailBench CrescendoXsafety_score | #11 / 21 | ↓21.38 | Source ↗official |