Qwen3.5 397B A17B vs Qwen3.6 Plus
Wins 2 of 8 areas
Reasoning · Following instructions
Wins 5 of 8 areas
Coding · Agents · Facts · Images and charts · Math
Qwen3.6 Plus is the stronger all-rounder.Qwen3.5 397B A17B is better at reasoning.
Scores updated · 56 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 software17Qwen3.6 Plus7 of 8 tests
- MathCompetition and research-level math03Qwen3.6 Plus3 of 4 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own02Qwen3.6 Plus2 of 2 tests
- FactsGetting facts right instead of making them up12Qwen3.6 Plus2 of 3 tests
- Images and chartsUnderstanding pictures, charts and video01Qwen3.6 Plus1 of 2 tests · 1 tie
- ReasoningHard problems that need careful thinking21Qwen3.5 397B A17B2 of 3 tests
- Following instructionsDoing exactly what it is asked10Qwen3.5 397B A17B1 of 1 test
- Long documentsFinding answers in very long texts11Even1 each
Following instructions 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 Plus costs 17% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Qwen3.6 Plus pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+54.3points ahead
- Short factual questions, answered correctlySimpleQA Verified+18.1points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+10.1points ahead
Where Qwen3.5 397B A17B pulls ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+4.4points ahead
- Code for real scientific research problemsSciCode+4.1points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+3.6points ahead
Every test, side by side
All 56 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3.6 Plus
- Terminal-Bench 2.1Qwen3.6 Plus by 10.151.361.4+10.1
- LMArena · WebDevQwen3.6 Plus by 61 rating points14001461+61 rating
- SWE-bench ProQwen3.6 Plus by 5.750.956.6+5.7
- SWE-bench MultilingualQwen3.6 Plus by 4.569.373.8+4.5
- SciCodeQwen3.5 397B A17B by 4.144.840.7+4.1
- LiveCodeBench v6Qwen3.6 Plus by 3.583.687.1+3.5
- Terminal-Bench HardQwen3.6 Plus by 340.943.9+3
- SWE-bench VerifiedQwen3.6 Plus by 2.476.478.8+2.4
AgentsQwen3.6 Plus
- GDPValQwen3.6 Plus by 9.914.824.7+9.9
- τ-Bench V3 · BankingQwen3.6 Plus by 7.413.420.8+7.4
ReasoningQwen3.5 397B A17B
- CritPtQwen3.6 Plus by 1.21.72.9+1.2
- Humanity's Last ExamQwen3.5 397B A17B by 1.22927.8+1.2
- GPQA DiamondQwen3.5 397B A17B by 1.189.388.2+1.1
FactsQwen3.6 Plus
- AA-Omniscience · Non-hallucinationQwen3.6 Plus by 54.311.165.4+54.3
- SimpleQA VerifiedQwen3.6 Plus by 18.12644.1+18.1
- AA-Omniscience · AccuracyQwen3.5 397B A17B by 4.430.826.4+4.4
Images and chartsQwen3.6 Plus
MathQwen3.6 Plus
- FrontierMath Tiers 1-3 (v2)Qwen3.6 Plus by 7.131.238.3+7.1
- AIME 2026Qwen3.6 Plus by 491.395.3+4
- IMOAnswerBenchQwen3.6 Plus by 2.980.983.8+2.9
- HMMT Feb. 2026tie87.987.8tie
Long documentsEven
- LongBench v2Qwen3.5 397B A17B by 1.263.262+1.2
- AA-LCRQwen3.6 Plus by 177.378.3+1
Following instructionsQwen3.5 397B A17B
- IFBenchQwen3.5 397B A17B by 3.678.875.2+3.6
Other results31 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-OmniscienceQwen3.6 Plus by 31.6-30.80.9+31.6
- AA Agentic IndexQwen3.6 Plus by 18.410.629+18.4
- AgentWorldBench - MCPQwen3.5 397B A17B by 1368.355.3+13
- Terminal-Bench 2.0Qwen3.6 Plus by 9.152.561.6+9.1
- AgentWorldBench - SearchQwen3.5 397B A17B by 8.930.821.9+8.9
- AA IntelligenceQwen3.6 Plus by 8.618.427+8.6
- DeepPlanningQwen3.6 Plus by 7.234.341.5+7.2
- Artificial Analysis Coding IndexQwen3.6 Plus by 6.348.254.5+6.3
- NL2RepoQwen3.6 Plus by 5.732.237.9+5.7
- VITA-BenchQwen3.5 397B A17B by 5.449.744.3+5.4
- AgentWorldBench - SWEQwen3.5 397B A17B by 5.364.459.1+5.3
- AgentWorldBench - TerminalQwen3.5 397B A17B by 4.755.350.6+4.7
- AgentWorldBench - OverallQwen3.5 397B A17B by 3.954.750.8+3.9
- AgentWorldBench - AndroidQwen3.6 Plus by 2.854.957.6+2.8
- HLE (with tools)Qwen3.6 Plus by 2.348.350.6+2.3
- AgentWorldBench - WebQwen3.6 Plus by 2.248.550.8+2.2
- MCPMarkQwen3.6 Plus by 2.146.148.2+2.1
- τ²-Bench Telecom (AA run)Qwen3.6 Plus by 2.195.697.7+2.1
- HMMT 2025Qwen3.6 Plus by 1.992.794.6+1.9
- HMMT Nov. 2025Qwen3.6 Plus by 1.992.794.6+1.9
- IFEvalQwen3.6 Plus by 1.792.694.3+1.7
- ToolathlonQwen3.6 Plus by 1.538.339.8+1.5
- SuperGPQAQwen3.6 Plus by 1.270.471.6+1.2
- MMMLUQwen3.6 Plus by 188.589.5+1
- MMLU-Protie87.888.5tie
- AgentWorldBench - OStie60.960.3tie
- Includetie85.685.1tie
- MMLU-Reduxtie94.994.5tie
- C-Evaltie9393.3tie
- WideSearchtie7474.3tie
- MMLU-ProXtie84.784.7tie
Questions people ask
Which is better, Qwen3.5 397B A17B or Qwen3.6 Plus?
Qwen3.6 Plus wins five of the eight areas we test: coding, agents, facts, images and charts and math. Qwen3.5 397B A17B wins reasoning and following instructions. They are level on long documents.
Which is better for coding?
Qwen3.6 Plus. It wins 7 of the 8 coding tests both models report; Qwen3.5 397B A17B wins 1.
Which is cheaper?
Qwen3.5 397B A17B costs $0.60 per million input tokens and $3.60 per million output tokens; Qwen3.6 Plus costs $0.50 and $3.00. That makes Qwen3.6 Plus about 17% cheaper for the same work.
How do you compare the two?
We use the 56 benchmark tests both models have published scores on. The verdict counts the 25 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 31 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.