Gemma 3n E4B

Gemma 3n E4B.

Google DeepMind Released 26 Jun 2025 Family Gemma 3n The lab's model card

Benchmarks with a score

Ranked first, out of five or more models

In the top three

Google DeepMind Jun 2025

Open weights Small

No Epoch Capabilities Index score yet

Where it ranks

Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.

Benchmark	Score	Rank	Source
BoolQ	81.6%	#14 of 28	Model card
ARC-Easy	81.6%	#14 of 23	Model card
LMArena (Elo)	1,318	#211 of 333	Official leaderboard
ARC-Challenge	61.6%	#32 of 45	Model card
PIQA	81%	#24 of 33	Model card
TriviaQA	70.2%	#29 of 39	Model card
HellaSwag	78.6%	#35 of 47	Model card
WinoGrande	71.7%	#33 of 42	Model card
Global MMLU Lite	64.5%	#13 of 16	Model card
SIQA	50%	#18 of 22	Model card
MMLU	64.9%	#107 of 127	Model card
Natural Questions	20.9%	#17 of 20	Model card
BIG-Bench Hard	52.9%	#61 of 71	Model card
MGSM	60.7%	#45 of 52	Model card
Wer wird Millionär? LLM Benchmark	111.0	#67 of 76	Official leaderboard
DROP	60.8%	#36 of 40	Model card
INCLUDE	57.2%	#23 of 25	Model card
MMLU-Pro	50.6%	#139 of 151	Model card
MMMLU	35.6%	#62 of 62	Model card
MMLU-ProX	19.9%	#17 of 17	Model card
Global MMLU	60.3%	#11 of 11	Model card

Benchmark	Score	Rank	Source
HiddenMath	37.7%	#9 of 11	Model card
DTBench	16%	#164 of 178	Official leaderboard
Misguided Attention	22.8%	#44 of 45	Official leaderboard
GPQA Diamond	23.7%	#283 of 284	Model card
AIME 2025	11.6%	#143 of 143	Model card

Benchmark	Score	Rank	Source
HumanEval	75%	#56 of 95	Model card
MBPP	63.6%	#48 of 66	Model card
LiveCodeBench	13.2%	#71 of 73	Model card
LiveCodeBench v5	25.7%	#43 of 44	Model card

Benchmark	Score		Rank	Source
Audio MC	15.5%		#14 of 16	Official leaderboard
MMAU v05.15.25	58%		#9 of 9	Official leaderboard

Benchmark	Score		Rank	Source
ACCoRD	41.2		#129 of 150	Official leaderboard
Conceptual Reasoning Index	13.3		#139 of 149	Official leaderboard

The benchmarks the lab chose to headline, with the scores it gave.