Gemma 4 E2B vs Gemma 4 E4B

Change models
IndexAll categories, out of 100.
21.9 ±5.5
28.2 ±5.2 best
GapIndex points below the top.
−6.3 · no clear gap
Top
CoverageIndex weight that has results.
90%
90%
SpeedOutput tokens per second.
—
—
EloLMArena rating and rank.
—
—
ReleasedMonth of release.
2026-04
2026-04
Input / 1MUS dollars per 1M input tokens.
—
—
Output / 1MUS dollars per 1M output tokens.
—
—
1K chats2K tokens in, 500 out, each.
—
—
1K documents20K tokens in, 1K out, each.
—
—
1K agent tasks200K tokens in, 20K out, each.
—
—
ContextMaximum tokens in one request.
128K
128K
AgenticMulti-step tasks with tools.
25.3
25.3
CodingCode writing and repair.
20.5
27.4 best
ReasoningLogic problems and puzzles.
25.6
30.9 best
MultimodalTasks with images and text.
16.8
32.8 best
KnowledgeFacts and expert knowledge.
22.1
30.2 best
MultilingualTasks in many languages.
5.0
8.8 best
InstructionTasks with strict prompt rules.
24.5
30.3 best
MathMath problems.
—
—

Benchmarks

22 shared
Only Gemma 4 E4B: 2
Gemma 4 E2B
Gemma 4 E4B
4.8% best181 / 222
3.8%202 / 222
16.3%195 / 217
32.0% best177 / 217
0.0%150 / 214
0.6% best132 / 214
EuroEval PolishMultilingual
20.2%173 / 209
25.2% best163 / 209
43.3%183 / 199
57.6% best162 / 199
EuroEval SwedishMultilingual
19.3%162 / 195
29.2% best134 / 195
EuroEval GermanMultilingual
12.9%173 / 191
21.7% best142 / 191
19.5%157 / 182
27.6% best140 / 182
EuroEval FrenchMultilingual
24.3%144 / 178
30.1% best135 / 178
EuroEval SpanishMultilingual
14.6%152 / 176
23.6% best131 / 176
EuroEval DutchMultilingual
24.1%146 / 175
33.2% best121 / 175
EuroEval ItalianMultilingual
16.6%148 / 174
26.4% best130 / 174
20.8%130 / 149
20.8%130 / 149
38.0%114 / 146
44.2% best88 / 146
43.4%129 / 141
58.6% best114 / 141
0.0%101 / 133
0.0%101 / 133
44.6%107 / 113
51.4% best102 / 113
60.0%64 / 75
69.4% best56 / 75
44.0%55 / 59
52.0% best51 / 59
44.2%52 / 53
52.6% best48 / 53