Gemma 4 E2B vs Ministral 3 3B
Wins 3 of 7 areas
Coding · Reasoning · Following instructions
Wins 0 of 7 areas
—
Gemma 4 E2B wins more areas, narrowly.
Scores updated · 29 tests both models report · How we compare
Where each one wins
Tests won in each of the seven areas where both have results. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- CodingWriting and fixing software20Gemma 4 E2B2 of 2 tests
- ReasoningHard problems that need careful thinking10Gemma 4 E2B1 of 3 tests · 2 ties
- Following instructionsDoing exactly what it is asked10Gemma 4 E2B1 of 1 test
- AgentsCarrying out multi-step tasks on its own00Even0 each · 2 ties
- FactsGetting facts right instead of making them up11Even1 each
- Images and chartsUnderstanding pictures, charts and video11Even1 each
- Long documentsFinding answers in very long texts00Even0 each · 1 tie
Long documents 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 E2B pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+47.8points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+11.2points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+7.5points ahead
Where Ministral 3 3B pulls ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+2.4points ahead
- Reads text in imagesOCRBench+1.6points ahead
Every test, side by side
All 29 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGemma 4 E2B
- SciCodeGemma 4 E2B by 5.620.915.3+5.6
- Terminal-Bench HardGemma 4 E2B by 330+3
ReasoningGemma 4 E2B
- GPQA DiamondGemma 4 E2B by 7.543.335.8+7.5
- Humanity's Last Examtie4.85.4tie
- CritPttie00tie
FactsEven
- AA-Omniscience · Non-hallucinationGemma 4 E2B by 47.867.619.8+47.8
- AA-Omniscience · AccuracyMinistral 3 3B by 2.46.69+2.4
Images and chartsEven
Following instructionsGemma 4 E2B
- IFBenchGemma 4 E2B by 11.23826.8+11.2
Other results16 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.
- AA-OmniscienceGemma 4 E2B by 40.4-23.6-64+40.4
- ChartQA TestMinistral 3 3B by 35.943.279.1+35.9
- GQA TestDev_BalancedMinistral 3 3B by 15.738.754.4+15.7
- HallusionBenchMinistral 3 3B by 5.459.865.2+5.4
- τ²-Bench Telecom (AA run)Ministral 3 3B by 4.120.824.9+4.1
- DocVQA-valMinistral 3 3B by 3.985.789.6+3.9
- MMLU-Pro testMinistral 3 3B by 3.444.147.5+3.4
- MMLU testGemma 4 E2B by 3.269.266+3.2
- MMMU DEV_VALMinistral 3 3B by 3.147.750.8+3.1
- AA IntelligenceGemma 4 E2B by 37.84.8+3
- MTL MMBench_DEVMinistral 3 3B by 2.664.867.4+2.6
- Artificial Analysis Coding IndexGemma 4 E2B by 2.47.24.8+2.4
- OCRBench v2_enGemma 4 E2B by 2.143.541.4+2.1
- MMMBtie74.373.4tie
- MMBench DEV_EN_V11tie69.369.2tie
- MMBench_DEVtie0.60.7tie
Questions people ask
Which is better, Gemma 4 E2B or Ministral 3 3B?
Gemma 4 E2B wins three of the seven areas where both have results: coding, reasoning and following instructions. Ministral 3 3B wins none. They are level on agents, facts, images and charts and long documents.
Which is better for coding?
Gemma 4 E2B. It wins 2 of the 2 coding tests both models report; Ministral 3 3B wins none.
How do you compare the two?
We use the 29 benchmark tests both models have published scores on. The verdict counts the 13 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 16 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.