Qwen3.5 27B vs Qwen3 Max (Reasoning)
Wins 2 of 7 areas
Agents · Long documents
Wins 2 of 7 areas
Reasoning · Math
The two are evenly matched.Qwen3.5 27B is better at long documents and agents; Qwen3 Max (Reasoning) at math and reasoning.
Scores updated · 33 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.
- MathCompetition and research-level math02Qwen3 Max (Reasoning)2 of 2 tests
- ReasoningHard problems that need careful thinking01Qwen3 Max (Reasoning)1 of 3 tests · 2 ties
- Long documentsFinding answers in very long texts10Qwen3.5 27B1 of 2 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own10Qwen3.5 27B1 of 1 test
- CodingWriting and fixing software22Even2 each
- FactsGetting facts right instead of making them up11Even1 each
- Following instructionsDoing exactly what it is asked11Even1 each
Agents 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.
Qwen3.5 27B costs 63% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Qwen3 Max (Reasoning) pulls ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+9.1points ahead
- Recent programming contest problemsLiveCodeBench v6+5.2points ahead
- Very hard expert questions across many subjectsHumanity's Last Exam+4.1points ahead
Where Qwen3.5 27B pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+13.2points ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+8.4points ahead
- Finds hard-to-locate facts by browsing the webBrowseComp+7.1points ahead
Every test, side by side
All 33 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingEven
- Terminal-Bench HardQwen3.5 27B by 8.432.624.2+8.4
- LiveCodeBench v6Qwen3 Max (Reasoning) by 5.280.785.9+5.2
- SWE-bench VerifiedQwen3 Max (Reasoning) by 2.972.475.3+2.9
- SWE-bench MultilingualQwen3.5 27B by 2.669.366.7+2.6
ReasoningQwen3 Max (Reasoning)
- Humanity's Last ExamQwen3 Max (Reasoning) by 4.123.928+4.1
- CritPttie0.91.7tie
- GPQA Diamondtie85.886.1tie
FactsEven
- AA-Omniscience · Non-hallucinationQwen3.5 27B by 13.218.55.3+13.2
- AA-Omniscience · AccuracyQwen3 Max (Reasoning) by 9.120.729.8+9.1
MathQwen3 Max (Reasoning)
- IMOAnswerBenchQwen3 Max (Reasoning) by 479.983.9+4
- AIME 2026Qwen3 Max (Reasoning) by 1.691.793.3+1.6
Long documentsQwen3.5 27B
- AA-LCRQwen3.5 27B by 3.477.774.3+3.4
- LongBench v2tie60.660.6tie
Following instructionsEven
- IFBenchQwen3.5 27B by 4.975.670.7+4.9
- Multi-ChallengeQwen3 Max (Reasoning) by 2.560.863.3+2.5
Other results17 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.
- Terminal-Bench 2.0Qwen3.5 27B by 19.141.622.5+19.1
- τ²-Bench Telecom (AA run)Qwen3.5 27B by 10.393.983.6+10.3
- AA-OmniscienceQwen3 Max (Reasoning) by 7.4-44-36.6+7.4
- DeepPlanningQwen3 Max (Reasoning) by 6.122.628.7+6.1
- t2-benchQwen3.5 27B by 4.27974.8+4.2
- C-EvalQwen3 Max (Reasoning) by 3.290.593.7+3.2
- WideSearchQwen3.5 27B by 3.261.157.9+3.2
- MCPMarkQwen3.5 27B by 2.836.333.5+2.8
- SuperGPQAQwen3 Max (Reasoning) by 1.765.667.3+1.7
- AA IntelligenceQwen3.5 27B by 1.622.921.3+1.6
- MMMLUQwen3.5 27B by 1.585.984.4+1.5
- BrowseComp-ZHQwen3.5 27B by 1.262.160.9+1.2
- VITA-BenchQwen3.5 27B by 141.940.9+1
- BFCLv4tie68.567.7tie
- MMLU-Protie86.185.7tie
- MMLU-Reduxtie93.292.8tie
- Seal-0tie47.246.9tie
Questions people ask
Which is better, Qwen3.5 27B or Qwen3 Max (Reasoning)?
Qwen3.5 27B and Qwen3 Max (Reasoning) each win two of the seven areas where both have results. Qwen3.5 27B wins agents and long documents; Qwen3 Max (Reasoning) wins reasoning and math. Qwen3.5 27B costs 63% less. They are level on coding, facts and following instructions.
Which is better for coding?
Neither. They win 2 coding tests each of the 4 both models report.
Which is cheaper?
Qwen3.5 27B costs $0.30 per million input tokens and $2.40 per million output tokens; Qwen3 Max (Reasoning) costs $1.20 and $6.00. That makes Qwen3.5 27B about 63% cheaper for the same work.
How do you compare the two?
We use the 33 benchmark tests both models have published scores on. The verdict counts the 16 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 17 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.