Gemma 4 E2B vs Qwen3.5 4B
Wins 0 of 8 areas
—
Wins 7 of 8 areas
Coding · Agents · Reasoning · Images and charts · Math · Long documents · Following instructions
Qwen3.5 4B is the stronger all-rounder.
Scores updated · 58 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 video05Qwen3.5 4B5 of 5 tests
- ReasoningHard problems that need careful thinking02Qwen3.5 4B2 of 3 tests · 1 tie
- CodingWriting and fixing software12Qwen3.5 4B2 of 3 tests
- AgentsCarrying out multi-step tasks on its own01Qwen3.5 4B1 of 1 test
- MathCompetition and research-level math01Qwen3.5 4B1 of 1 test
- Long documentsFinding answers in very long texts01Qwen3.5 4B1 of 1 test
- Following instructionsDoing exactly what it is asked01Qwen3.5 4B1 of 1 test
- FactsGetting facts right instead of making them up11Even1 each
Agents, math, 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 Qwen3.5 4B pulls ahead
- US invitational high-school math exam problemsAIME 2026+52.2points ahead
- Reasons across sets of long documentsAA-LCR+46.7points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+33.8points ahead
Where Gemma 4 E2B pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+54.2points ahead
- Code for real scientific research problemsSciCode+4.8points ahead
Every test, side by side
All 58 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3.5 4B
- Terminal-Bench HardQwen3.5 4B by 15.2318.2+15.2
- LiveCodeBench v6Qwen3.5 4B by 11.84455.8+11.8
- SciCodeGemma 4 E2B by 4.820.916.1+4.8
ReasoningQwen3.5 4B
- GPQA DiamondQwen3.5 4B by 33.843.377.1+33.8
- Humanity's Last ExamQwen3.5 4B by 5.14.89.9+5.1
- CritPttie00tie
FactsEven
- AA-Omniscience · Non-hallucinationGemma 4 E2B by 54.267.613.4+54.2
- AA-Omniscience · AccuracyQwen3.5 4B by 8.56.615.1+8.5
Images and chartsQwen3.5 4B
Other results41 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 4B by 71.320.892.1+71.3
- Tau² TelecomQwen3.5 4B by 65.322.487.7+65.3
- ScreenSpot-v2 WebQwen3.5 4B by 55.422.477.8+55.4
- τ²-Bench (Retail)Qwen3.5 4B by 52.918.971.9+52.9
- t2-benchQwen3.5 4B by 50.529.479.9+50.5
- ScreenSpot-v2 DesktopQwen3.5 4B by 48.228.176.3+48.2
- ChartQA TestQwen3.5 4B by 4143.284.2+41
- ScreenSpot-v2 MobileQwen3.5 4B by 38.542.981.4+38.5
- AA-OmniscienceGemma 4 E2B by 34.7-23.6-58.4+34.7
- AA Agentic IndexQwen3.5 4B by 311.532.5+31
- PinchBenchQwen3.5 4B by 27.144.271.3+27.1
- InfographicVQA (val)Qwen3.5 4B by 25.954.480.3+25.9
- MMEQwen3.5 4B by 24.654.979.5+24.6
- AIME25 no toolsQwen3.5 4B by 2326.349.3+23
- MMLU-ProQwen3.5 4B by 19.16079.1+19.1
- TextVQA-valQwen3.5 4B by 18.762.581.2+18.7
- MATH500 (Pass@1)Qwen3.5 4B by 16.86480.8+16.8
- BrowseComp+ (OpenClaw)Qwen3.5 4B by 16.28.324.5+16.2
- Artificial Analysis Coding IndexQwen3.5 4B by 15.47.222.6+15.4
- OCRBench v1Qwen3.5 4B by 15.470.285.6+15.4
- OCRBench v2_enQwen3.5 4B by 15.243.558.7+15.2
- BFCL v3Qwen3.5 4B by 14.756.471.1+14.7
- MMBenchQwen3.5 4B by 14.264.278.4+14.2
- Multilingual MMBenchQwen3.5 4B by 14.262.877+14.2
- BFCLv4Qwen3.5 4B by 13.63750.6+13.6
- SimpleVQAQwen3.5 4B by 13.427.340.7+13.4
- Claw-Eval AvgQwen3.5 4B by 9.253.162.3+9.2
- MMMU (val) (Pass@1)Qwen3.5 4B by 9.241.150.3+9.2
- DocVQA-valQwen3.5 4B by 9.185.794.8+9.1
- MMMLUQwen3.5 4B by 8.767.476.1+8.7
- Multilingual MMMBQwen3.5 4B by 8.773.382+8.7
- HallusionBenchGemma 4 E2B by 8.159.851.7+8.1
- MMMBQwen3.5 4B by 7.774.382+7.7
- RealWorldQAQwen3.5 4B by 7.16067.1+7.1
- IFEvalQwen3.5 4B by 6.88389.8+6.8
- LiveCodeBenchV6 no toolsQwen3.5 4B by 654.960.9+6
- AA IntelligenceQwen3.5 4B by 5.37.813.1+5.3
- SEED-Bench (image)Qwen3.5 4B by 4.771.476.1+4.7
- MM-IFEvalGemma 4 E2B by 2.565.663.1+2.5
- τ³-Bench BankingQwen3.5 4B by 2.13.45.5+2.1
- POPEQwen3.5 4B by 28486+2
Questions people ask
Which is better, Gemma 4 E2B or Qwen3.5 4B?
Qwen3.5 4B wins seven of the eight areas we test: coding, agents, reasoning, images and charts, math, long documents and following instructions. Gemma 4 E2B wins none. They are level on facts.
Which is better for coding?
Qwen3.5 4B. It wins 2 of the 3 coding tests both models report; Gemma 4 E2B wins 1.
How do you compare the two?
We use the 58 benchmark tests both models have published scores on. The verdict counts the 17 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 41 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.