Gemma 4 E4B vs Qwen3.5 2B
Wins 6 of 7 areas
Coding · Agents · Reasoning · Facts · Long documents · Following instructions
Wins 0 of 7 areas
—
Gemma 4 E4B is the stronger all-rounder.
Scores updated · 46 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.
- ReasoningHard problems that need careful thinking20Gemma 4 E4B2 of 3 tests · 1 tie
- CodingWriting and fixing software21Gemma 4 E4B2 of 3 tests
- AgentsCarrying out multi-step tasks on its own10Gemma 4 E4B1 of 2 tests · 1 tie
- FactsGetting facts right instead of making them up10Gemma 4 E4B1 of 2 tests · 1 tie
- Long documentsFinding answers in very long texts10Gemma 4 E4B1 of 1 test
- Following instructionsDoing exactly what it is asked10Gemma 4 E4B1 of 1 test
- Images and chartsUnderstanding pictures, charts and video22Even2 each
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 three tests each model wins by the widest margin. Scores are out of 100.
Where Gemma 4 E4B pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+43.3points ahead
- Code for real scientific research problemsSciCode+21.6points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+12.7points ahead
Where Qwen3.5 2B pulls ahead
- Math problems shown in pictures and chartsMathVista+3.5points ahead
- Questions that truly need the image to answerMMStar+2.2points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+1.1points ahead
Every test, side by side
All 46 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGemma 4 E4B
- SciCodeGemma 4 E4B by 21.624.42.8+21.6
- Terminal-Bench HardGemma 4 E4B by 4.58.33.8+4.5
- Terminal-Bench 2.1Qwen3.5 2B by 1.11.93+1.1
ReasoningGemma 4 E4B
- GPQA DiamondGemma 4 E4B by 1257.645.6+12
- Humanity's Last ExamGemma 4 E4B by 1.23.82.6+1.2
- CritPttie0.60tie
FactsGemma 4 E4B
- AA-Omniscience · Non-hallucinationGemma 4 E4B by 43.369.125.8+43.3
- AA-Omniscience · Accuracytie8.68.1tie
Images and chartsEven
Following instructionsGemma 4 E4B
- IFBenchGemma 4 E4B by 12.744.231.5+12.7
Other results30 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.
- τ²-Bench Telecom (AA run)Qwen3.5 2B by 48.220.869+48.2
- AA-OmniscienceGemma 4 E4B by 40.5-19.7-60.2+40.5
- ChartQA TestQwen3.5 2B by 36.342.178.4+36.3
- ScreenSpot-v2 WebQwen3.5 2B by 18.347.665.9+18.3
- ScreenSpot-v2 DesktopQwen3.5 2B by 1845.863.8+18
- HallusionBenchQwen3.5 2B by 15.749.865.5+15.7
- IFEvalGemma 4 E4B by 14.387.973.6+14.3
- MMMLUGemma 4 E4B by 13.576.663.1+13.5
- MM-IFEvalGemma 4 E4B by 12.868.255.4+12.8
- InfographicVQA (val)Qwen3.5 2B by 12.660.973.5+12.6
- OCRBench v1Qwen3.5 2B by 10.973.584.4+10.9
- ScreenSpot-v2 MobileQwen3.5 2B by 9.460.369.7+9.4
- t2-benchGemma 4 E4B by 8.757.548.8+8.7
- MMEQwen3.5 2B by 8.667.676.2+8.6
- TextVQA-valQwen3.5 2B by 8.36977.3+8.3
- Artificial Analysis Coding IndexGemma 4 E4B by 6.59.42.9+6.5
- BFCLv4Gemma 4 E4B by 6.14033.9+6.1
- MMMBGemma 4 E4B by 5.980.474.5+5.9
- DocVQA-valQwen3.5 2B by 5.287.492.6+5.2
- MMMU (val) (Pass@1)Gemma 4 E4B by 5.249.344.1+5.2
- SimpleVQAQwen3.5 2B by 4.830.435.2+4.8
- Multilingual MMMBGemma 4 E4B by 4.580.475.9+4.5
- MMLU-ProGemma 4 E4B by 2.969.466.5+2.9
- AA IntelligenceGemma 4 E4B by 28.96.9+2
- POPEQwen3.5 2B by 1.786.988.6+1.7
- MMBenchQwen3.5 2B by 1.571.673.1+1.5
- Multilingual MMBenchGemma 4 E4B by 1.371.269.9+1.3
- RealWorldQAtie64.365.1tie
- OCRBench v2_entie48.848.1tie
- SEED-Bench (image)tie75.375.8tie
Questions people ask
Which is better, Gemma 4 E4B or Qwen3.5 2B?
Gemma 4 E4B wins six of the seven areas where both have results: coding, agents, reasoning, facts, long documents and following instructions. Qwen3.5 2B wins none. They are level on images and charts.
Which is better for coding?
Gemma 4 E4B. It wins 2 of the 3 coding tests both models report; Qwen3.5 2B wins 1.
How do you compare the two?
We use the 46 benchmark tests both models have published scores on. The verdict counts the 16 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 30 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.