Qwen3.5 35B A3B vs Qwen3.6 27B
Wins 1 of 8 areas
Following instructions
Wins 7 of 8 areas
Coding · Agents · Reasoning · Facts · Images and charts · Math · Long documents
Qwen3.6 27B is the stronger all-rounder.Qwen3.5 35B A3B is cheaper and better at following instructions.
Both rank among the ten best models we track in images and charts.
Scores updated · 56 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.
- CodingWriting and fixing software07Qwen3.6 27B7 of 7 tests
- Images and chartsUnderstanding pictures, charts and video14Qwen3.6 27B4 of 7 tests · 2 ties
- MathCompetition and research-level math02Qwen3.6 27B2 of 3 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own02Qwen3.6 27B2 of 2 tests
- ReasoningHard problems that need careful thinking01Qwen3.6 27B1 of 3 tests · 2 ties
- FactsGetting facts right instead of making them up01Qwen3.6 27B1 of 2 tests · 1 tie
- Long documentsFinding answers in very long texts01Qwen3.6 27B1 of 1 test
- Following instructionsDoing exactly what it is asked10Qwen3.5 35B A3B1 of 1 test
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.
Qwen3.5 35B A3B costs 46% 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 Qwen3.6 27B pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+36.1points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+19.9points ahead
- Real work tasks from 44 professionsGDPVal+18.8points ahead
Every test, side by side
All 56 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3.6 27B
- Terminal-Bench 2.1Qwen3.6 27B by 19.940.860.7+19.9
- SWE-bench MultilingualQwen3.6 27B by 1160.371.3+11
- LiveCodeBench v6Qwen3.6 27B by 9.374.683.9+9.3
- SWE-bench ProQwen3.6 27B by 8.944.653.5+8.9
- Terminal-Bench HardQwen3.6 27B by 8.326.534.8+8.3
- SWE-bench VerifiedQwen3.6 27B by 869.277.2+8
- SciCodeQwen3.6 27B by 5.137.742.8+5.1
AgentsQwen3.6 27B
- GDPValQwen3.6 27B by 18.85.624.4+18.8
- τ-Bench V3 · BankingQwen3.6 27B by 11.84.916.7+11.8
ReasoningQwen3.6 27B
- Humanity's Last ExamQwen3.6 27B by 2.12123.1+2.1
- GPQA Diamondtie84.584.2tie
- CritPttie0.91.1tie
FactsQwen3.6 27B
- AA-Omniscience · Non-hallucinationQwen3.6 27B by 36.114.650.7+36.1
- AA-Omniscience · Accuracytie20.119.6tie
Images and chartsQwen3.6 27B
- OCRBenchv2Qwen3.6 27B by 365.368.3+3
- MMMU-ProQwen3.6 27B by 1.972.774.6+1.9
- OCRBenchQwen3.5 35B A3B by 1.69189.4+1.6
- MMMUQwen3.6 27B by 1.581.482.9+1.5
- MathVistaQwen3.6 27B by 1.286.287.4+1.2
- CharXiv (RQ)tie77.578.4tie
- MMStartie81.981.4tie
MathQwen3.6 27B
- IMOAnswerBenchQwen3.6 27B by 476.880.8+4
- HMMT Feb. 2026Qwen3.6 27B by 2.581.884.3+2.5
- AIME 2026tie93.394.1tie
Following instructionsQwen3.5 35B A3B
- IFBenchQwen3.5 35B A3B by 4.972.567.6+4.9
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.
- QwenWebBenchQwen3.6 27B by 509 rating points9781487+509 rating
- SkillsBench Avg5Qwen3.6 27B by 43.84.448.2+43.8
- AA-OmniscienceQwen3.6 27B by 28.1-48.1-20+28.1
- Terminal-Bench 2.0Qwen3.6 27B by 18.840.559.3+18.8
- Artificial Analysis Coding IndexQwen3.6 27B by 16.73753.7+16.7
- NL2RepoQwen3.6 27B by 15.720.536.2+15.7
- Claw Eval (pass@3)Qwen3.6 27B by 9.65160.6+9.6
- AA Agentic IndexQwen3.6 27B by 8.311.820.1+8.3
- Claw-Eval AvgQwen3.6 27B by 765.472.4+7
- RefSpatial-BenchQwen3.6 27B by 6.563.570+6.5
- QwenClawBenchQwen3.6 27B by 5.747.753.4+5.7
- τ²-Bench Telecom (AA run)Qwen3.6 27B by 589.294.2+5
- HMMT Feb. 2025Qwen3.6 27B by 4.88993.8+4.8
- VideoMMMUQwen3.6 27B by 480.484.4+4
- OCRBench v2 (Chinese)Qwen3.5 35B A3B by 361.758.7+3
- SuperGPQAQwen3.6 27B by 2.663.466+2.6
- ERQAQwen3.5 35B A3B by 2.364.862.5+2.3
- SimpleVQAQwen3.5 35B A3B by 2.258.356.1+2.2
- AA IntelligenceQwen3.6 27B by 2.119.321.4+2.1
- EmbSpatial-BenchQwen3.6 27B by 1.583.184.6+1.5
- HMMT 2025Qwen3.6 27B by 1.589.290.7+1.5
- HMMT Nov. 2025Qwen3.6 27B by 1.589.290.7+1.5
- C-EvalQwen3.6 27B by 1.290.291.4+1.2
- MLVUQwen3.6 27B by 185.686.6+1
- MMLU-Protie85.386.2tie
- MV-Benchtie74.875.5tie
- DynaMathtie8585.6tie
- CC-OCRtie80.781.2tie
- MMLU-Reduxtie93.393.5tie
- RealWorldQAtie84.184.1tie
Questions people ask
Which is better, Qwen3.5 35B A3B or Qwen3.6 27B?
Qwen3.6 27B wins seven of the eight areas we test: coding, agents, reasoning, facts, images and charts, math and long documents. Qwen3.5 35B A3B wins following instructions, and costs 46% less.
Which is better for coding?
Qwen3.6 27B. It wins 7 of the 7 coding tests both models report; Qwen3.5 35B A3B wins none.
Which is cheaper?
Qwen3.5 35B A3B costs $0.25 per million input tokens and $2.00 per million output tokens; Qwen3.6 27B costs $0.60 and $3.60. That makes Qwen3.5 35B A3B about 46% cheaper for the same work.
How do you compare the two?
We use the 56 benchmark tests both models have published scores on. The verdict counts the 26 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.