Gemma 4 E2B vs Gemma 4 E4B
Wins 0 of 8 areas
—
Wins 6 of 8 areas
Coding · Facts · Images and charts · Math · Long documents · Following instructions
Gemma 4 E4B is the stronger all-rounder.
Scores updated · 67 tests both models report · How we compare
Where each one wins
Tests won in each of the eight areas we test. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- Images and chartsUnderstanding pictures, charts and video04Gemma 4 E4B4 of 4 tests
- CodingWriting and fixing software03Gemma 4 E4B3 of 3 tests
- FactsGetting facts right instead of making them up02Gemma 4 E4B2 of 2 tests
- Long documentsFinding answers in very long texts02Gemma 4 E4B2 of 2 tests
- MathCompetition and research-level math01Gemma 4 E4B1 of 1 test
- Following instructionsDoing exactly what it is asked01Gemma 4 E4B1 of 1 test
- AgentsCarrying out multi-step tasks on its own00Even0 each · 2 ties
- ReasoningHard problems that need careful thinking11Even1 each · 1 tie
Math and following instructions rest on a single test each.
What it costs
Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where Gemma 4 E4B pulls ahead
- Reasons across sets of long documentsAA-LCR+15.7points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+14.3points ahead
- Recent programming contest problemsLiveCodeBench v6+8points ahead
Where Gemma 4 E2B pulls ahead
- Very hard expert questions across many subjectsHumanity's Last Exam+1points ahead
Every test, side by side
All 67 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGemma 4 E4B
- LiveCodeBench v6Gemma 4 E4B by 84452+8
- Terminal-Bench HardGemma 4 E4B by 5.338.3+5.3
- SciCodeGemma 4 E4B by 3.520.924.4+3.5
ReasoningEven
- GPQA DiamondGemma 4 E4B by 14.343.357.6+14.3
- Humanity's Last ExamGemma 4 E2B by 14.83.8+1
- CritPttie00.6tie
FactsGemma 4 E4B
- AA-Omniscience · AccuracyGemma 4 E4B by 26.68.6+2
- AA-Omniscience · Non-hallucinationGemma 4 E4B by 1.567.669.1+1.5
Images and chartsGemma 4 E4B
Long documentsGemma 4 E4B
- AA-LCRGemma 4 E4B by 15.716.332+15.7
- MRCR v2 8 needle 128k (average)Gemma 4 E4B by 6.319.125.4+6.3
Following instructionsGemma 4 E4B
- IFBenchGemma 4 E4B by 6.23844.2+6.2
Other results49 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- Codeforces ELOGemma 4 E4B by 307 rating points633940+307 rating
- t2-benchGemma 4 E4B by 28.129.457.5+28.1
- ScreenSpot-v2 WebGemma 4 E4B by 25.222.447.6+25.2
- τ²-Bench (Retail)Gemma 4 E4B by 23.118.942.1+23.1
- ScreenSpot-v2 DesktopGemma 4 E4B by 17.728.145.8+17.7
- τ²-BenchGemma 4 E4B by 17.724.542.2+17.7
- ScreenSpot-v2 MobileGemma 4 E4B by 17.442.960.3+17.4
- MMEGemma 4 E4B by 12.754.967.6+12.7
- BIG-Bench Extra HardGemma 4 E4B by 11.221.933.1+11.2
- PinchBenchGemma 4 E4B by 10.944.255.1+10.9
- HallusionBenchGemma 4 E2B by 1059.849.8+10
- MMLU-ProGemma 4 E4B by 9.46069.4+9.4
- MMMLUGemma 4 E4B by 9.267.476.6+9.2
- LiveCodeBenchV6 no toolsGemma 4 E4B by 8.954.963.8+8.9
- MMBenchGemma 4 E4B by 8.462.871.2+8.4
- Multilingual MMBenchGemma 4 E4B by 8.462.871.2+8.4
- MMMU (val) (Pass@1)Gemma 4 E4B by 8.241.149.3+8.2
- AIME25 no toolsGemma 4 E4B by 826.334.3+8
- BrowseComp+ (OpenClaw)Gemma 4 E4B by 7.68.315.9+7.6
- MATH-VisionGemma 4 E4B by 7.152.459.5+7.1
- Multilingual MMMBGemma 4 E4B by 7.173.380.4+7.1
- BFCLv4Gemma 4 E4B by 6.833.240+6.8
- InfographicVQA (val)Gemma 4 E4B by 6.554.460.9+6.5
- TextVQA-valGemma 4 E4B by 6.562.569+6.5
- MMMBGemma 4 E4B by 6.174.380.4+6.1
- OCRBench v2_enGemma 4 E4B by 5.343.548.8+5.3
- MedXPertQA MMGemma 4 E4B by 5.223.528.7+5.2
- Claw-Eval AvgGemma 4 E4B by 4.953.158+4.9
- IFEvalGemma 4 E4B by 4.98387.9+4.9
- Tau² TelecomGemma 4 E4B by 4.422.426.8+4.4
- RealWorldQAGemma 4 E4B by 4.36064.3+4.3
- AA-OmniscienceGemma 4 E4B by 3.9-23.6-19.7+3.9
- SEED-Bench (image)Gemma 4 E4B by 3.971.475.3+3.9
- OCRBench v1Gemma 4 E4B by 3.370.273.5+3.3
- SimpleVQAGemma 4 E4B by 3.127.330.4+3.1
- POPEGemma 4 E4B by 2.98486.9+2.9
- MM-IFEvalGemma 4 E4B by 2.665.668.2+2.6
- Artificial Analysis Coding IndexGemma 4 E4B by 2.27.29.4+2.2
- CoVoSTGemma 4 E4B by 233.535.5+2
- DocVQA-valGemma 4 E4B by 1.785.787.4+1.7
- AA IntelligenceGemma 4 E4B by 1.17.88.9+1.1
- ChartQA TestGemma 4 E2B by 1.143.242.1+1.1
- MATH500 (Pass@1)Gemma 4 E4B by 16465+1
- BFCL v3tie56.457.3tie
- τ³-Bench Bankingtie3.44.1tie
- AA Agentic Indextie1.51.8tie
- OmniDocBench 1.5 (average edit distance, lower is better)lower is bettertie0.30.2tie
- FLEURS (lower is better)lower is bettertie0.10.1tie
- τ²-Bench Telecom (AA run)tie20.820.8tie
Questions people ask
Which is better, Gemma 4 E2B or Gemma 4 E4B?
Gemma 4 E4B wins six of the eight areas we test: coding, facts, images and charts, math, long documents and following instructions. Gemma 4 E2B wins none. They are level on agents and reasoning.
Which is better for coding?
Gemma 4 E4B. It wins 3 of the 3 coding tests both models report; Gemma 4 E2B wins none.
How do you compare the two?
We use the 67 benchmark tests both models have published scores on. The verdict counts the 18 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 49 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.