Gemma 3 4B vs Qwen3 235B A22B
Wins 1 of 6 areas
Long documents
Wins 5 of 6 areas
Coding · Agents · Reasoning · Facts · Following instructions
Qwen3 235B A22B is the stronger all-rounder.Gemma 3 4B is cheaper and better at long documents.
Scores updated · 32 tests both models report · How we compare
Where each one wins
Tests won in each of the six 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 software03Qwen3 235B A22B3 of 3 tests
- ReasoningHard problems that need careful thinking02Qwen3 235B A22B2 of 3 tests · 1 tie
- FactsGetting facts right instead of making them up12Qwen3 235B A22B2 of 3 tests
- AgentsCarrying out multi-step tasks on its own01Qwen3 235B A22B1 of 1 test
- Following instructionsDoing exactly what it is asked01Qwen3 235B A22B1 of 1 test
- Long documentsFinding answers in very long texts10Gemma 3 4B1 of 1 test
Agents, 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.
Gemma 3 4B costs 96% less for the same work.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where Qwen3 235B A22B pulls ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+40.9points ahead
- Code for real scientific research problemsSciCode+32.6points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+20.6points ahead
Where Gemma 3 4B pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 32 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3 235B A22B
- SciCodeQwen3 235B A22B by 32.67.339.9+32.6
- LiveCodeBench v6Qwen3 235B A22B by 18.118.937+18.1
- Terminal-Bench HardQwen3 235B A22B by 5.30.86.1+5.3
ReasoningQwen3 235B A22B
- GPQA DiamondQwen3 235B A22B by 40.929.170+40.9
- Humanity's Last ExamQwen3 235B A22B by 5.75.311+5.7
- CritPttie00tie
FactsQwen3 235B A22B
- AA-Omniscience · Non-hallucinationQwen3 235B A22B by 20.61.822.4+20.6
- AA-Omniscience · AccuracyQwen3 235B A22B by 10.87.718.5+10.8
- Vectara HHEM hallucination ratelower is betterGemma 3 4B by 2.96.49.3+2.9
Following instructionsQwen3 235B A22B
- IFBenchQwen3 235B A22B by 10.428.338.7+10.4
Other results20 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.
- Arena HardQwen3 235B A22B by 63.831.895.6+63.8
- LiveCodeBenchQwen3 235B A22B by 58.112.670.7+58.1
- AA-OmniscienceQwen3 235B A22B by 38.2-82.9-44.7+38.2
- MMMLUQwen3 235B A22B by 36.650.186.7+36.6
- MMLUQwen3 235B A22B by 28.658.487+28.6
- vectara_answer_rateQwen3 235B A22B by 27.667.394.9+27.6
- MMLU-ProQwen3 235B A22B by 24.643.668.2+24.6
- τ²-Bench Telecom (AA run)Qwen3 235B A22B by 19524+19
- MBPPQwen3 235B A22B by 18.263.281.4+18.2
- MATH-500 (EM)Qwen3 235B A22B by 1873.291.2+18
- AA Agentic IndexQwen3 235B A22B by 16.71.718.4+16.7
- BBHQwen3 235B A22B by 16.772.288.9+16.7
- Artificial Analysis Coding IndexQwen3 235B A22B by 14.72.717.4+14.7
- SimpleQAQwen3 235B A22B by 9.2413.2+9.2
- IFEvalGemma 3 4B by 790.283.2+7
- GSM8KQwen3 235B A22B by 5.289.294.4+5.2
- AA IntelligenceQwen3 235B A22B by 4.74.89.5+4.7
- MGSMGemma 3 4B by 3.887.383.5+3.8
- vectara_factual_consistencyGemma 3 4B by 2.993.690.7+2.9
- vectara_avg_summary_lengthQwen3 235B A22B by 28.2 rating points77.4105.6+28.2 rating
Questions people ask
Which is better, Gemma 3 4B or Qwen3 235B A22B?
Qwen3 235B A22B wins five of the six areas where both have results: coding, agents, reasoning, facts and following instructions. Gemma 3 4B wins long documents, and costs 96% less.
Which is better for coding?
Qwen3 235B A22B. It wins 3 of the 3 coding tests both models report; Gemma 3 4B wins none.
Which is cheaper?
Gemma 3 4B costs $0.05 per million input tokens and $0.10 per million output tokens; Qwen3 235B A22B costs $0.70 and $2.80. That makes Gemma 3 4B about 96% cheaper for the same work.
How do you compare the two?
We use the 32 benchmark tests both models have published scores on. The verdict counts the 12 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 20 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.