Gemini 3 Pro vs Qwen3.5 27B
Wins 4 of 8 areas
Coding · Reasoning · Images and charts · Math
Wins 2 of 8 areas
Agents · Facts
Gemini 3 Pro wins more areas, narrowly.Qwen3.5 27B is cheaper and better at agents.
Both rank among the ten best models we track in images and charts.
Scores updated · 52 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 video40Gemini 3 Pro4 of 5 tests · 1 tie
- CodingWriting and fixing software52Gemini 3 Pro5 of 7 tests
- ReasoningHard problems that need careful thinking30Gemini 3 Pro3 of 3 tests
- MathCompetition and research-level math20Gemini 3 Pro2 of 3 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own12Qwen3.5 27B2 of 3 tests
- FactsGetting facts right instead of making them up12Qwen3.5 27B2 of 3 tests
- Long documentsFinding answers in very long texts11Even1 each
- Following instructionsDoing exactly what it is asked11Even1 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.
Qwen3.5 27B costs 81% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Gemini 3 Pro pulls ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+35.1points ahead
- Code for real scientific research problemsSciCode+16.6points ahead
- Very hard expert questions across many subjectsHumanity's Last Exam+15.8points ahead
Where Qwen3.5 27B pulls ahead
- Multi-step tasks using many real software toolsMCP Atlas+14.3points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+10points ahead
- Long, multi-file coding tasks in real codebasesSWE-bench Pro+7.9points ahead
Every test, side by side
All 52 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGemini 3 Pro
- SciCodeGemini 3 Pro by 16.656.139.5+16.6
- Terminal-Bench HardGemini 3 Pro by 9.141.732.6+9.1
- LMArena · WebDevGemini 3 Pro by 83 rating points14401357+83 rating
- SWE-bench ProQwen3.5 27B by 7.943.351.2+7.9
- LiveCodeBench v6Gemini 3 Pro by 6.787.480.7+6.7
- SWE-bench MultilingualQwen3.5 27B by 4.36569.3+4.3
- SWE-bench VerifiedGemini 3 Pro by 3.876.272.4+3.8
AgentsQwen3.5 27B
- MCP AtlasQwen3.5 27B by 14.354.168.4+14.3
- BrowseCompQwen3.5 27B by 1.859.261+1.8
- GDPValGemini 3 Pro by 1.234.233+1.2
ReasoningGemini 3 Pro
- Humanity's Last ExamGemini 3 Pro by 15.839.723.9+15.8
- CritPtGemini 3 Pro by 8.29.10.9+8.2
- GPQA DiamondGemini 3 Pro by 590.885.8+5
FactsQwen3.5 27B
- AA-Omniscience · AccuracyGemini 3 Pro by 35.155.820.7+35.1
- AA-Omniscience · Non-hallucinationQwen3.5 27B by 108.518.5+10
- Vectara HHEM hallucination ratelower is betterQwen3.5 27B by 1.513.612.1+1.5
Images and chartsGemini 3 Pro
- LMArena · VisionGemini 3 Pro by 65 rating points13051240+65 rating
- MMMU-ProGemini 3 Pro by 5.280.275+5.2
- MathVistaGemini 3 Pro by 289.887.8+2
- CharXiv (RQ)Gemini 3 Pro by 1.981.479.5+1.9
- OCRBenchtie90.389.4tie
MathGemini 3 Pro
- HMMT Feb. 2026Gemini 3 Pro by 5.386.481.1+5.3
- IMOAnswerBenchGemini 3 Pro by 3.283.179.9+3.2
- AIME 2026tie91.791.7tie
Long documentsEven
- LongBench v2Gemini 3 Pro by 7.668.260.6+7.6
- AA-LCRQwen3.5 27B by 1.77677.7+1.7
Following instructionsEven
- IFBenchQwen3.5 27B by 5.270.475.6+5.2
- Multi-ChallengeGemini 3 Pro by 4.965.760.8+4.9
Other results24 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-OmniscienceGemini 3 Pro by 59.315.3-44+59.3
- SimpleVQAGemini 3 Pro by 13.769.756+13.7
- Terminal-Bench 2.0Gemini 3 Pro by 12.654.241.6+12.6
- Artificial Analysis Coding IndexGemini 3 Pro by 11.646.534.9+11.6
- τ²-Bench Telecom (AA run)Qwen3.5 27B by 6.887.193.9+6.8
- t2-benchGemini 3 Pro by 6.485.479+6.4
- MMMLUGemini 3 Pro by 5.991.885.9+5.9
- HMMT Feb. 2025Gemini 3 Pro by 5.597.592+5.5
- VideoMMMUGemini 3 Pro by 5.387.682.3+5.3
- MLVUQwen3.5 27B by 5.280.785.9+5.2
- AA IntelligenceGemini 3 Pro by 5.12822.9+5.1
- MMVUGemini 3 Pro by 4.277.573.3+4.2
- MMLU-ProGemini 3 Pro by 490.186.1+4
- HMMT Nov. 2025Gemini 3 Pro by 3.593.389.8+3.5
- AA Agentic IndexQwen3.5 27B by 2.65254.6+2.6
- ScreenSpot-Pro (No tools)Gemini 3 Pro by 2.472.770.3+2.4
- ZeroBenchQwen3.5 27B by 2810+2
- Seal-0Qwen3.5 27B by 1.745.547.2+1.7
- vectara_factual_consistencyQwen3.5 27B by 1.586.487.9+1.5
- vectara_avg_summary_lengthtie101.994.4tie
- OmniDocBench 1.5tie88.588.9tie
- vectara_answer_ratetie99.499.8tie
- LVBenchtie73.573.6tie
- MATH-Visiontie86.186tie
Questions people ask
Which is better, Gemini 3 Pro or Qwen3.5 27B?
Gemini 3 Pro wins four of the eight areas we test: coding, reasoning, images and charts and math. Qwen3.5 27B wins agents and facts, and costs 81% less. They are level on long documents and following instructions.
Which is better for coding?
Gemini 3 Pro. It wins 5 of the 7 coding tests both models report; Qwen3.5 27B wins 2.
Which is cheaper?
Gemini 3 Pro costs $2.00 per million input tokens and $12.00 per million output tokens; Qwen3.5 27B costs $0.30 and $2.40. That makes Qwen3.5 27B about 81% cheaper for the same work.
How do you compare the two?
We use the 52 benchmark tests both models have published scores on. The verdict counts the 28 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 24 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.