Qwen3.5 397B A17B vs Qwen3.6 35B A3B
Wins 7 of 8 areas
Coding · Agents · Reasoning · Images and charts · Math · Long documents · Following instructions
Wins 0 of 8 areas
—
Qwen3.5 397B A17B is the stronger all-rounder.Qwen3.6 35B A3B is cheaper.
Scores updated · 51 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 software70Qwen3.5 397B A17B7 of 7 tests
- ReasoningHard problems that need careful thinking30Qwen3.5 397B A17B3 of 3 tests
- MathCompetition and research-level math31Qwen3.5 397B A17B3 of 4 tests
- Images and chartsUnderstanding pictures, charts and video20Qwen3.5 397B A17B2 of 2 tests
- AgentsCarrying out multi-step tasks on its own21Qwen3.5 397B A17B2 of 4 tests · 1 tie
- Long documentsFinding answers in very long texts10Qwen3.5 397B A17B1 of 1 test
- Following instructionsDoing exactly what it is asked10Qwen3.5 397B A17B1 of 1 test
- FactsGetting facts right instead of making them up11Even1 each
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.
Qwen3.6 35B A3B costs 38% 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
- Finds hard-to-locate facts by browsing the webBrowseComp+20.3points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+14.4points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+12points ahead
Where Qwen3.6 35B A3B pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+38.4points ahead
- Real work tasks from 44 professionsGDPVal+5points ahead
- US invitational high-school math exam problemsAIME 2026+1.4points ahead
Every test, side by side
All 51 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3.5 397B A17B
- SciCodeQwen3.5 397B A17B by 8.244.836.6+8.2
- Terminal-Bench 2.1Qwen3.5 397B A17B by 6.451.344.9+6.4
- Terminal-Bench HardQwen3.5 397B A17B by 6.140.934.8+6.1
- LiveCodeBench v6Qwen3.5 397B A17B by 3.283.680.4+3.2
- SWE-bench VerifiedQwen3.5 397B A17B by 376.473.4+3
- SWE-bench MultilingualQwen3.5 397B A17B by 2.169.367.2+2.1
- SWE-bench ProQwen3.5 397B A17B by 1.450.949.5+1.4
AgentsQwen3.5 397B A17B
- BrowseCompQwen3.5 397B A17B by 20.36948.7+20.3
- GDPValQwen3.6 35B A3B by 514.819.8+5
- τ-Bench V3 · BankingQwen3.5 397B A17B by 4.113.49.3+4.1
- Terminal-Bench 4.0tie00tie
ReasoningQwen3.5 397B A17B
- Humanity's Last ExamQwen3.5 397B A17B by 6.82922.2+6.8
- GPQA DiamondQwen3.5 397B A17B by 5.289.384.1+5.2
- CritPtQwen3.5 397B A17B by 1.41.70.3+1.4
FactsEven
- AA-Omniscience · Non-hallucinationQwen3.6 35B A3B by 38.411.149.5+38.4
- AA-Omniscience · AccuracyQwen3.5 397B A17B by 1230.818.8+12
Images and chartsQwen3.5 397B A17B
MathQwen3.5 397B A17B
- FrontierMath Tiers 1-3 (v2)Qwen3.5 397B A17B by 10.831.220.4+10.8
- HMMT Feb. 2026Qwen3.5 397B A17B by 4.387.983.6+4.3
- IMOAnswerBenchQwen3.5 397B A17B by 280.978.9+2
- AIME 2026Qwen3.6 35B A3B by 1.491.392.7+1.4
Long documentsQwen3.5 397B A17B
- AA-LCRQwen3.5 397B A17B by 5.677.371.7+5.6
Following instructionsQwen3.5 397B A17B
- IFBenchQwen3.5 397B A17B by 14.478.864.4+14.4
Other results27 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 35B A3B by 211 rating points11861397+211 rating
- VITA-BenchQwen3.5 397B A17B by 14.149.735.6+14.1
- WideSearchQwen3.5 397B A17B by 13.97460.1+13.9
- ToolathlonQwen3.5 397B A17B by 11.438.326.9+11.4
- MCPMarkQwen3.5 397B A17B by 9.146.137+9.1
- AA-OmniscienceQwen3.6 35B A3B by 8.5-30.8-22.2+8.5
- DeepPlanningQwen3.5 397B A17B by 8.434.325.9+8.4
- Artificial Analysis Coding IndexQwen3.5 397B A17B by 6.348.241.9+6.3
- SuperGPQAQwen3.5 397B A17B by 5.770.464.7+5.7
- GDPval-AA v2Qwen3.6 35B A3B by 53 rating points9621015+53 rating
- AA Agentic IndexQwen3.6 35B A3B by 4.410.615+4.4
- HMMT Feb. 2025Qwen3.5 397B A17B by 4.194.890.7+4.1
- HMMT 2025Qwen3.5 397B A17B by 3.692.789.1+3.6
- HMMT Nov. 2025Qwen3.5 397B A17B by 3.692.789.1+3.6
- C-EvalQwen3.5 397B A17B by 39390+3
- τ³-Bench BankingQwen3.5 397B A17B by 2.913.410.5+2.9
- NL2RepoQwen3.5 397B A17B by 2.832.229.4+2.8
- MMLU-ProQwen3.5 397B A17B by 2.687.885.2+2.6
- Claw-Eval AvgQwen3.5 397B A17B by 270.768.7+2
- MMLU-ReduxQwen3.5 397B A17B by 1.694.993.3+1.6
- PinchBenchQwen3.6 35B A3B by 1.586.688.1+1.5
- SkillsBench Avg5Qwen3.5 397B A17B by 1.33028.7+1.3
- GPQA (unspecified)Qwen3.5 397B A17B by 1.187.186+1.1
- Terminal-Bench 2.0Qwen3.5 397B A17B by 152.551.5+1
- QwenClawBenchtie51.852.6tie
- τ²-Bench Telecom (AA run)tie95.695.3tie
- AA Intelligencetie18.418.2tie
Questions people ask
Which is better, Qwen3.5 397B A17B or Qwen3.6 35B A3B?
Qwen3.5 397B A17B wins seven of the eight areas we test: coding, agents, reasoning, images and charts, math, long documents and following instructions. Qwen3.6 35B A3B wins none, but costs 38% less. They are level on facts.
Which is better for coding?
Qwen3.5 397B A17B. It wins 7 of the 7 coding tests both models report; Qwen3.6 35B A3B wins none.
Which is cheaper?
Qwen3.5 397B A17B costs $0.60 per million input tokens and $3.60 per million output tokens; Qwen3.6 35B A3B costs $0.38 and $2.25. That makes Qwen3.6 35B A3B about 38% cheaper for the same work.
How do you compare the two?
We use the 51 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 27 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.