Qwen3.5 397B A17B vs Qwen3 Max (Reasoning)
Wins 6 of 7 areas
Coding · Agents · Reasoning · Facts · Long documents · Following instructions
Wins 1 of 7 areas
Math
Qwen3.5 397B A17B is the stronger all-rounder.Qwen3 Max (Reasoning) is better at math.
Scores updated · 34 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 software31Qwen3.5 397B A17B3 of 4 tests
- ReasoningHard problems that need careful thinking20Qwen3.5 397B A17B2 of 3 tests · 1 tie
- FactsGetting facts right instead of making them up20Qwen3.5 397B A17B2 of 2 tests
- Long documentsFinding answers in very long texts20Qwen3.5 397B A17B2 of 2 tests
- Following instructionsDoing exactly what it is asked20Qwen3.5 397B A17B2 of 2 tests
- AgentsCarrying out multi-step tasks on its own10Qwen3.5 397B A17B1 of 1 test
- MathCompetition and research-level math02Qwen3 Max (Reasoning)2 of 2 tests
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 397B A17B costs 42% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Qwen3.5 397B A17B pulls ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+16.7points ahead
- Finds hard-to-locate facts by browsing the webBrowseComp+15.1points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+8.1points ahead
Where Qwen3 Max (Reasoning) pulls ahead
- Olympiad math problems with short answersIMOAnswerBench+3points ahead
- Recent programming contest problemsLiveCodeBench v6+2.3points ahead
- US invitational high-school math exam problemsAIME 2026+2points ahead
Every test, side by side
All 34 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3.5 397B A17B
- Terminal-Bench HardQwen3.5 397B A17B by 16.740.924.2+16.7
- SWE-bench MultilingualQwen3.5 397B A17B by 2.669.366.7+2.6
- LiveCodeBench v6Qwen3 Max (Reasoning) by 2.383.685.9+2.3
- SWE-bench VerifiedQwen3.5 397B A17B by 1.176.475.3+1.1
ReasoningQwen3.5 397B A17B
- GPQA DiamondQwen3.5 397B A17B by 3.289.386.1+3.2
- Humanity's Last ExamQwen3.5 397B A17B by 12928+1
- CritPttie1.71.7tie
FactsQwen3.5 397B A17B
- AA-Omniscience · Non-hallucinationQwen3.5 397B A17B by 5.811.15.3+5.8
- AA-Omniscience · AccuracyQwen3.5 397B A17B by 130.829.8+1
MathQwen3 Max (Reasoning)
- IMOAnswerBenchQwen3 Max (Reasoning) by 380.983.9+3
- AIME 2026Qwen3 Max (Reasoning) by 291.393.3+2
Long documentsQwen3.5 397B A17B
- AA-LCRQwen3.5 397B A17B by 377.374.3+3
- LongBench v2Qwen3.5 397B A17B by 2.663.260.6+2.6
Following instructionsQwen3.5 397B A17B
- IFBenchQwen3.5 397B A17B by 8.178.870.7+8.1
- Multi-ChallengeQwen3.5 397B A17B by 4.367.663.3+4.3
Other results18 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 397B A17B by 3052.522.5+30
- ToolathlonQwen3.5 397B A17B by 19.538.318.8+19.5
- WideSearchQwen3.5 397B A17B by 16.17457.9+16.1
- MCPMarkQwen3.5 397B A17B by 12.646.133.5+12.6
- τ²-Bench Telecom (AA run)Qwen3.5 397B A17B by 1295.683.6+12
- t2-benchQwen3.5 397B A17B by 11.986.774.8+11.9
- BrowseComp-ZHQwen3.5 397B A17B by 9.470.360.9+9.4
- VITA-BenchQwen3.5 397B A17B by 8.849.740.9+8.8
- AA-OmniscienceQwen3.5 397B A17B by 5.9-30.8-36.6+5.9
- DeepPlanningQwen3.5 397B A17B by 5.634.328.7+5.6
- BFCLv4Qwen3.5 397B A17B by 5.272.967.7+5.2
- MMMLUQwen3.5 397B A17B by 4.188.584.4+4.1
- SuperGPQAQwen3.5 397B A17B by 3.170.467.3+3.1
- AA IntelligenceQwen3 Max (Reasoning) by 2.918.421.3+2.9
- MMLU-ProQwen3.5 397B A17B by 2.187.885.7+2.1
- MMLU-ReduxQwen3.5 397B A17B by 2.194.992.8+2.1
- C-Evaltie9393.7tie
- Seal-0tie46.946.9tie
Questions people ask
Which is better, Qwen3.5 397B A17B or Qwen3 Max (Reasoning)?
Qwen3.5 397B A17B wins six of the seven areas where both have results: coding, agents, reasoning, facts, long documents and following instructions. Qwen3 Max (Reasoning) wins math.
Which is better for coding?
Qwen3.5 397B A17B. It wins 3 of the 4 coding tests both models report; Qwen3 Max (Reasoning) wins 1.
Which is cheaper?
Qwen3.5 397B A17B costs $0.60 per million input tokens and $3.60 per million output tokens; Qwen3 Max (Reasoning) costs $1.20 and $6.00. That makes Qwen3.5 397B A17B about 42% cheaper for the same work.
How do you compare the two?
We use the 34 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 18 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.