Qwen3.6 27B vs Qwen3.7 Max
Wins 0 of 7 areas
—
Wins 7 of 7 areas
Coding · Agents · Reasoning · Facts · Math · Long documents · Following instructions
Qwen3.7 Max is the stronger all-rounder.Qwen3.6 27B is cheaper.
Scores updated · 42 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.
- CodingWriting and fixing software09Qwen3.7 Max9 of 9 tests
- MathCompetition and research-level math05Qwen3.7 Max5 of 5 tests
- ReasoningHard problems that need careful thinking04Qwen3.7 Max4 of 4 tests
- FactsGetting facts right instead of making them up02Qwen3.7 Max2 of 2 tests
- Following instructionsDoing exactly what it is asked02Qwen3.7 Max2 of 2 tests
- AgentsCarrying out multi-step tasks on its own12Qwen3.7 Max2 of 3 tests
- Long documentsFinding answers in very long texts01Qwen3.7 Max1 of 1 test
Long documents rests on a single test.
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.
Qwen3.6 27B costs 58% 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 Qwen3.7 Max pulls ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+29.5points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+23.7points ahead
- Follows detailed instructions exactlyLiveBench · Instruction Following+20.8points ahead
Where Qwen3.6 27B pulls ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+4.9points ahead
Every test, side by side
All 42 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3.7 Max
- Terminal-Bench HardQwen3.7 Max by 1634.850.8+16
- Terminal-Bench 2.1Qwen3.7 Max by 13.860.774.5+13.8
- LiveCodeBench v6Qwen3.7 Max by 7.783.991.6+7.7
- SWE-bench ProQwen3.7 Max by 7.153.560.6+7.1
- SWE-bench MultilingualQwen3.7 Max by 771.378.3+7
- SciCodeQwen3.7 Max by 6.742.849.5+6.7
- LiveBench · Agentic CodingQwen3.7 Max by 4.339.343.6+4.3
- SWE-bench VerifiedQwen3.7 Max by 3.277.280.4+3.2
- LiveBench · CodingQwen3.7 Max by 2.471.874.2+2.4
AgentsQwen3.7 Max
- GDPValQwen3.7 Max by 7.224.431.6+7.2
- τ-Bench V3 · BankingQwen3.6 27B by 4.916.711.8+4.9
- Terminal-Bench 4.0Qwen3.7 Max by 1.501.5+1.5
ReasoningQwen3.7 Max
- Humanity's Last ExamQwen3.7 Max by 17.423.140.5+17.4
- LiveBench · ReasoningQwen3.7 Max by 1370.383.3+13
- CritPtQwen3.7 Max by 12.31.113.4+12.3
- GPQA DiamondQwen3.7 Max by 8.184.292.3+8.1
FactsQwen3.7 Max
- AA-Omniscience · Non-hallucinationQwen3.7 Max by 23.750.774.4+23.7
- AA-Omniscience · AccuracyQwen3.7 Max by 11.519.631.1+11.5
MathQwen3.7 Max
- FrontierMath Tiers 1-3 (v2)Qwen3.7 Max by 29.535.164.6+29.5
- HMMT Feb. 2026Qwen3.7 Max by 12.884.397.1+12.8
- IMOAnswerBenchQwen3.7 Max by 9.280.890+9.2
- LiveBench · MathematicsQwen3.7 Max by 5.479.985.3+5.4
- AIME 2026Qwen3.7 Max by 2.994.197+2.9
Following instructionsQwen3.7 Max
- LiveBench · Instruction FollowingQwen3.7 Max by 20.853.274+20.8
- IFBenchQwen3.7 Max by 12.967.680.5+12.9
Other results16 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-OmniscienceQwen3.7 Max by 33.5-2013.5+33.5
- livebench_languageQwen3.7 Max by 16.463.379.7+16.4
- Artificial Analysis Coding IndexQwen3.7 Max by 12.353.766+12.3
- NL2RepoQwen3.7 Max by 1136.247.2+11
- Terminal-Bench 2.0Qwen3.7 Max by 10.459.369.7+10.4
- JobBenchQwen3.7 Max by 9.521.831.3+9.5
- DeepSWE 1.1Qwen3.7 Max by 8.313.321.6+8.3
- AA IntelligenceQwen3.7 Max by 8.121.429.5+8.1
- SuperGPQAQwen3.7 Max by 7.66673.6+7.6
- Claw Eval (pass@3)Qwen3.7 Max by 4.660.665.2+4.6
- HMMT Nov. 2025Qwen3.7 Max by 4.390.795+4.3
- AA Agentic IndexQwen3.7 Max by 3.820.123.9+3.8
- MMLU-ProQwen3.7 Max by 3.486.289.6+3.4
- MMLU-ReduxQwen3.7 Max by 1.593.595+1.5
- livebench_data_analysisQwen3.7 Max by 1.470.471.8+1.4
- τ²-Bench Telecom (AA run)tie94.294.7tie
Questions people ask
Which is better, Qwen3.6 27B or Qwen3.7 Max?
Qwen3.7 Max wins all seven areas where both have results: coding, agents, reasoning, facts, math, long documents and following instructions. Qwen3.6 27B wins none, but costs 58% less.
Which is better for coding?
Qwen3.7 Max. It wins 9 of the 9 coding tests both models report; Qwen3.6 27B wins none.
Which is cheaper?
Qwen3.6 27B costs $0.60 per million input tokens and $3.60 per million output tokens; Qwen3.7 Max costs $2.50 and $7.50. That makes Qwen3.6 27B about 58% cheaper for the same work.
How do you compare the two?
We use the 42 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 16 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.