← Models

Model profile

Gemma 4 E2B

Googledeveloper
2026-04-02release date
Not rankedSafety rank
#233 / 645Freedom rank

Evidence summary

Safety. Gemma 4 E2B does not meet the evidence gate for a Safety rank. Published Safety evidence spans 2 eval lineages and 4 of 7 components. Its strongest relative result is AA-Omniscience (hallucination_rate, #39 of 345); its weakest is Enkrypt AI Safety Leaderboard (bias_attack_non_success_rate, #246 of 248).

Freedom. Gemma 4 E2B has an estimated Freedom rank of #233; its 90% source-sensitivity interval is #160–#380. Published Freedom evidence spans 1 eval lineages and 1 of 1 components. Its strongest relative result is Enkrypt AI Safety Leaderboard (cbrn_attack_non_success_rate, #19 of 248); its weakest is Enkrypt AI Safety Leaderboard (harmful_attack_non_success_rate, #180 of 248).

Compare this model

Only models sharing at least one published sub-eval are listed.

Official and reference links

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-evalRankValueDistributionSource
AA-Omnisciencehallucination_rate#39 / 3450.3236Source ↗official
Enkrypt AI Safety Leaderboardbias_attack_non_success_rate#246 / 2481.55Source ↗official
Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate#230 / 24868Source ↗official
Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate#67 / 24886.67Source ↗official
Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate#122 / 24696.36Source ↗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-evalRankValueDistributionSource
Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate#19 / 24868Source ↗official
Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate#180 / 24886.67Source ↗official
Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate#121 / 24696.36Source ↗official