Qwen3.5 397B A17B 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.Qwen3.5 397B A17B is cheaper.
Both rank among the ten best models we track in following instructions.
Scores updated · 45 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 software08Qwen3.7 Max8 of 8 tests
- MathCompetition and research-level math04Qwen3.7 Max4 of 4 tests
- ReasoningHard problems that need careful thinking03Qwen3.7 Max3 of 3 tests
- AgentsCarrying out multi-step tasks on its own13Qwen3.7 Max3 of 4 tests
- FactsGetting facts right instead of making them up02Qwen3.7 Max2 of 3 tests · 1 tie
- Long documentsFinding answers in very long texts01Qwen3.7 Max1 of 1 test
- Following instructionsDoing exactly what it is asked01Qwen3.7 Max1 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 397B A17B 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
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+63.3points ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+33.4points ahead
- Short factual questions, answered correctlySimpleQA Verified+29.8points ahead
Where Qwen3.5 397B A17B pulls ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+1.6points ahead
Every test, side by side
All 45 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3.7 Max
- Terminal-Bench 2.1Qwen3.7 Max by 23.251.374.5+23.2
- LMArena · WebDevQwen3.7 Max by 115 rating points14001515+115 rating
- Terminal-Bench HardQwen3.7 Max by 9.940.950.8+9.9
- SWE-bench ProQwen3.7 Max by 9.750.960.6+9.7
- SWE-bench MultilingualQwen3.7 Max by 969.378.3+9
- LiveCodeBench v6Qwen3.7 Max by 883.691.6+8
- SciCodeQwen3.7 Max by 4.744.849.5+4.7
- SWE-bench VerifiedQwen3.7 Max by 476.480.4+4
AgentsQwen3.7 Max
- GDPValQwen3.7 Max by 16.814.831.6+16.8
- AA IT-Bench SREQwen3.7 Max by 8.434.142.5+8.4
- τ-Bench V3 · BankingQwen3.5 397B A17B by 1.613.411.8+1.6
- Terminal-Bench 4.0Qwen3.7 Max by 1.501.5+1.5
ReasoningQwen3.7 Max
- CritPtQwen3.7 Max by 11.71.713.4+11.7
- Humanity's Last ExamQwen3.7 Max by 11.52940.5+11.5
- GPQA DiamondQwen3.7 Max by 389.392.3+3
FactsQwen3.7 Max
- AA-Omniscience · Non-hallucinationQwen3.7 Max by 63.311.174.4+63.3
- SimpleQA VerifiedQwen3.7 Max by 29.82655.8+29.8
- AA-Omniscience · Accuracytie30.831.1tie
MathQwen3.7 Max
- FrontierMath Tiers 1-3 (v2)Qwen3.7 Max by 33.431.264.6+33.4
- HMMT Feb. 2026Qwen3.7 Max by 9.287.997.1+9.2
- IMOAnswerBenchQwen3.7 Max by 9.180.990+9.1
- AIME 2026Qwen3.7 Max by 5.791.397+5.7
Following instructionsQwen3.7 Max
- IFBenchQwen3.7 Max by 1.778.880.5+1.7
Other results21 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 44.2-30.813.5+44.2
- Artificial Analysis Coding IndexQwen3.7 Max by 17.848.266+17.8
- Terminal-Bench 2.0Qwen3.7 Max by 17.252.569.7+17.2
- NL2RepoQwen3.7 Max by 1532.247.2+15
- MCPMarkQwen3.7 Max by 14.746.160.8+14.7
- AA Agentic IndexQwen3.7 Max by 13.310.623.9+13.3
- AA IntelligenceQwen3.7 Max by 11.118.429.5+11.1
- Toolathlon VerifiedQwen3.7 Max by 940.749.7+9
- HLE (with tools)Qwen3.7 Max by 5.248.353.5+5.2
- SuperGPQAQwen3.7 Max by 3.270.473.6+3.2
- HMMT Nov. 2025Qwen3.7 Max by 2.392.795+2.3
- MMLU-ProXQwen3.7 Max by 2.384.787+2.3
- BFCLv4Qwen3.7 Max by 2.172.975+2.1
- MMLU-ProQwen3.7 Max by 1.887.889.6+1.8
- MMMLUQwen3.7 Max by 1.888.590.3+1.8
- VITA-BenchQwen3.5 397B A17B by 1.849.747.9+1.8
- IFEvalQwen3.7 Max by 1.792.694.3+1.7
- WideSearchQwen3.7 Max by 1.27475.2+1.2
- τ²-Bench Telecom (AA run)tie95.694.7tie
- Includetie85.686.2tie
- MMLU-Reduxtie94.995tie
Questions people ask
Which is better, Qwen3.5 397B A17B 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. Qwen3.5 397B A17B wins none, but costs 58% less.
Which is better for coding?
Qwen3.7 Max. It wins 8 of the 8 coding tests both models report; Qwen3.5 397B A17B wins none.
Which is cheaper?
Qwen3.5 397B A17B 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 Qwen3.5 397B A17B about 58% cheaper for the same work.
How do you compare the two?
We use the 45 benchmark tests both models have published scores on. The verdict counts the 24 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 21 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.