Claude Sonnet 4.6 vs Qwen3.5 397B A17B
Wins 6 of 8 areas
Coding · Agents · Reasoning · Facts · Math · Long documents
Wins 1 of 8 areas
Following instructions
Claude Sonnet 4.6 is the stronger all-rounder.Qwen3.5 397B A17B is cheaper and better at following instructions.
Scores updated · 43 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 software60Claude Sonnet 4.66 of 6 tests
- AgentsCarrying out multi-step tasks on its own60Claude Sonnet 4.66 of 6 tests
- FactsGetting facts right instead of making them up30Claude Sonnet 4.63 of 3 tests
- ReasoningHard problems that need careful thinking21Claude Sonnet 4.62 of 3 tests
- MathCompetition and research-level math10Claude Sonnet 4.61 of 1 test
- Long documentsFinding answers in very long texts10Claude Sonnet 4.61 of 1 test
- Following instructionsDoing exactly what it is asked01Qwen3.5 397B A17B1 of 1 test
- Images and chartsUnderstanding pictures, charts and video11Even1 each
Math, 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 77% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Claude Sonnet 4.6 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+40.5points ahead
- Real work tasks from 44 professionsGDPVal+21.9points ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+21points ahead
Where Qwen3.5 397B A17B pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+22.2points ahead
- Harder college exam questions with imagesMMMU-Pro+4points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+1.8points ahead
Every test, side by side
All 43 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Sonnet 4.6
- Terminal-Bench 2.1Claude Sonnet 4.6 by 19.971.251.3+19.9
- LMArena · WebDevClaude Sonnet 4.6 by 122 rating points15221400+122 rating
- Terminal-Bench HardClaude Sonnet 4.6 by 12.15340.9+12.1
- SWE-bench ProClaude Sonnet 4.6 by 7.258.150.9+7.2
- SciCodeClaude Sonnet 4.6 by 5.350.144.8+5.3
- SWE-bench VerifiedClaude Sonnet 4.6 by 3.279.676.4+3.2
AgentsClaude Sonnet 4.6
- GDPValClaude Sonnet 4.6 by 21.936.714.8+21.9
- τ-Bench V3 · BankingClaude Sonnet 4.6 by 2134.413.4+21
- AA ApexAgentsClaude Sonnet 4.6 by 12.72815.3+12.7
- AA IT-Bench SREClaude Sonnet 4.6 by 5.739.834.1+5.7
- BrowseCompClaude Sonnet 4.6 by 5.774.769+5.7
- Terminal-Bench 4.0Claude Sonnet 4.6 by 330+3
ReasoningClaude Sonnet 4.6
- Humanity's Last ExamClaude Sonnet 4.6 by 4.633.629+4.6
- GPQA DiamondQwen3.5 397B A17B by 1.887.589.3+1.8
- CritPtClaude Sonnet 4.6 by 1.43.11.7+1.4
FactsClaude Sonnet 4.6
- AA-Omniscience · Non-hallucinationClaude Sonnet 4.6 by 40.551.611.1+40.5
- AA-Omniscience · AccuracyClaude Sonnet 4.6 by 10.140.930.8+10.1
- SimpleQA VerifiedClaude Sonnet 4.6 by 6.832.826+6.8
Images and chartsEven
- MMMU-ProQwen3.5 397B A17B by 473.377.3+4
- LMArena · VisionClaude Sonnet 4.6 by 19 rating points12831264+19 rating
Following instructionsQwen3.5 397B A17B
- IFBenchQwen3.5 397B A17B by 22.256.678.8+22.2
Other results20 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.
- GDPval-AA v2Claude Sonnet 4.6 by 433 rating points1395962+433 rating
- AA-OmniscienceClaude Sonnet 4.6 by 42.912.2-30.8+42.9
- AA Agentic IndexClaude Sonnet 4.6 by 22.533.110.6+22.5
- τ²-Bench Telecom (AA run)Qwen3.5 397B A17B by 19.975.795.6+19.9
- Artificial Analysis Coding IndexClaude Sonnet 4.6 by 14.86348.2+14.8
- AA IntelligenceClaude Sonnet 4.6 by 11.730.118.4+11.7
- Terminal-Bench 2.0Claude Sonnet 4.6 by 6.659.152.5+6.6
- AgentWorldBench - AndroidClaude Sonnet 4.6 by 3.15854.9+3.1
- GPQA (unspecified)Claude Sonnet 4.6 by 2.889.987.1+2.8
- ToolathlonClaude Sonnet 4.6 by 2.74138.3+2.7
- AgentWorldBench - OSClaude Sonnet 4.6 by 2.363.260.9+2.3
- AgentWorldBench - WebClaude Sonnet 4.6 by 2.250.848.5+2.2
- AgentWorldBench - SearchQwen3.5 397B A17B by 228.830.8+2
- AgentWorldBench - MCPClaude Sonnet 4.6 by 1.77068.3+1.7
- AgentWorldBench - TerminalClaude Sonnet 4.6 by 1.75755.3+1.7
- FLTEvalQwen3.5 397B A17B by 1.723.725.4+1.7
- HLE (with tools)Qwen3.5 397B A17B by 1.546.848.3+1.5
- AgentWorldBench - OverallClaude Sonnet 4.6 by 1.35654.7+1.3
- MMMLUtie89.388.5tie
- AgentWorldBench - SWEtie64.564.4tie
Questions people ask
Which is better, Claude Sonnet 4.6 or Qwen3.5 397B A17B?
Claude Sonnet 4.6 wins six of the eight areas we test: coding, agents, reasoning, facts, math and long documents. Qwen3.5 397B A17B wins following instructions, and costs 77% less. They are level on images and charts.
Which is better for coding?
Claude Sonnet 4.6. It wins 6 of the 6 coding tests both models report; Qwen3.5 397B A17B wins none.
Which is cheaper?
Claude Sonnet 4.6 costs $3.00 per million input tokens and $15.00 per million output tokens; Qwen3.5 397B A17B costs $0.60 and $3.60. That makes Qwen3.5 397B A17B about 77% cheaper for the same work.
How do you compare the two?
We use the 43 benchmark tests both models have published scores on. The verdict counts the 23 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 20 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.