Claude Opus 4.8 vs Qwen3.6 Plus
Wins 6 of 8 areas
Coding · Agents · Reasoning · Facts · Images and charts · Math
Wins 0 of 8 areas
—
Claude Opus 4.8 is the stronger all-rounder.Qwen3.6 Plus is cheaper.
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 software90Claude Opus 4.89 of 9 tests
- MathCompetition and research-level math40Claude Opus 4.84 of 5 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own40Claude Opus 4.84 of 4 tests
- ReasoningHard problems that need careful thinking40Claude Opus 4.84 of 4 tests
- FactsGetting facts right instead of making them up21Claude Opus 4.82 of 3 tests
- Images and chartsUnderstanding pictures, charts and video10Claude Opus 4.81 of 1 test
- Long documentsFinding answers in very long texts00Even0 each · 1 tie
- Following instructionsDoing exactly what it is asked11Even1 each
Images and charts and long documents 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 Plus costs 88% 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 Claude Opus 4.8 pulls ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+41.7points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+23.2points ahead
- Real work tasks from 44 professionsGDPVal+23.1points ahead
Where Qwen3.6 Plus pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+13points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+4.7points 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.
CodingClaude Opus 4.8
- Terminal-Bench 2.1Claude Opus 4.8 by 23.284.661.4+23.2
- Terminal-Bench HardClaude Opus 4.8 by 14.458.343.9+14.4
- SciCodeClaude Opus 4.8 by 13.754.440.7+13.7
- SWE-bench ProClaude Opus 4.8 by 12.669.256.6+12.6
- SWE-bench MultilingualClaude Opus 4.8 by 10.684.473.8+10.6
- SWE-bench VerifiedClaude Opus 4.8 by 9.888.678.8+9.8
- LMArena · WebDevClaude Opus 4.8 by 95 rating points15561461+95 rating
- LiveBench · Agentic CodingClaude Opus 4.8 by 9.150.541.4+9.1
- LiveBench · CodingClaude Opus 4.8 by 3.681.878.2+3.6
AgentsClaude Opus 4.8
- GDPValClaude Opus 4.8 by 23.147.824.7+23.1
- OSWorld-VerifiedClaude Opus 4.8 by 20.983.462.5+20.9
- τ-Bench V3 · BankingClaude Opus 4.8 by 13.434.220.8+13.4
- MCP AtlasClaude Opus 4.8 by 8.182.274.1+8.1
ReasoningClaude Opus 4.8
- Humanity's Last ExamClaude Opus 4.8 by 20.948.727.8+20.9
- CritPtClaude Opus 4.8 by 1820.92.9+18
- LiveBench · ReasoningClaude Opus 4.8 by 13.489.275.8+13.4
- GPQA DiamondClaude Opus 4.8 by 3.89288.2+3.8
FactsClaude Opus 4.8
- AA-Omniscience · AccuracyClaude Opus 4.8 by 22.448.826.4+22.4
- SimpleQA VerifiedClaude Opus 4.8 by 8.95344.1+8.9
- AA-Omniscience · Non-hallucinationQwen3.6 Plus by 4.760.765.4+4.7
Images and chartsClaude Opus 4.8
- CharXiv (RQ)Claude Opus 4.8 by 8.489.981.5+8.4
MathClaude Opus 4.8
- FrontierMath Tiers 1-3 (v2)Claude Opus 4.8 by 41.78038.3+41.7
- LiveBench · MathematicsClaude Opus 4.8 by 10.694.383.7+10.6
- HMMT Feb. 2026Claude Opus 4.8 by 7.795.587.8+7.7
- AIME 2026Claude Opus 4.8 by 4.710095.3+4.7
- IMOAnswerBenchtie83.583.8tie
Following instructionsEven
- LiveBench · Instruction FollowingClaude Opus 4.8 by 13.77258.3+13.7
- IFBenchQwen3.6 Plus by 1362.275.2+13
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.
- NL2RepoClaude Opus 4.8 by 31.869.737.9+31.8
- AA-OmniscienceClaude Opus 4.8 by 27.928.80.9+27.9
- frontiermath_tier_4_v1Claude Opus 4.8 by 2331.38.3+23
- Tool-DecathlonClaude Opus 4.8 by 20.159.939.8+20.1
- ToolathlonClaude Opus 4.8 by 20.159.939.8+20.1
- Artificial Analysis Coding IndexClaude Opus 4.8 by 19.874.354.5+19.8
- ScreenSpot-Pro (No tools)Claude Opus 4.8 by 19.787.968.2+19.7
- AA IntelligenceClaude Opus 4.8 by 14.841.827+14.8
- AA Agentic IndexClaude Opus 4.8 by 13.642.629+13.6
- AgentWorldBench - SearchClaude Opus 4.8 by 13.235.121.9+13.2
- Finance Agent v2Claude Opus 4.8 by 1353.940.9+13
- Terminal-Bench 2.0Claude Opus 4.8 by 1374.661.6+13
- AgentWorldBench - TerminalClaude Opus 4.8 by 8.659.250.6+8.6
- HLE (with tools)Claude Opus 4.8 by 7.357.950.6+7.3
- AgentWorldBench - OSClaude Opus 4.8 by 6.366.660.3+6.3
- LiveBenchClaude Opus 4.8 by 6.377.270.8+6.3
- AgentWorldBench - OverallClaude Opus 4.8 by 5.856.650.8+5.8
- AgentWorldBench - SWEClaude Opus 4.8 by 564.159.1+5
- livebench_languageClaude Opus 4.8 by 4.779.775+4.7
- AgentWorldBench - WebClaude Opus 4.8 by 3.954.750.8+3.9
- livebench_data_analysisQwen3.6 Plus by 3.96669.9+3.9
- AgentWorldBench - AndroidClaude Opus 4.8 by 3.861.557.6+3.8
- τ²-Bench Telecom (AA run)Qwen3.6 Plus by 3.394.497.7+3.3
- IncludeClaude Opus 4.8 by 2.587.685.1+2.5
- HMMT Nov. 2025Claude Opus 4.8 by 1.996.594.6+1.9
- WideSearchQwen3.6 Plus by 1.472.974.3+1.4
- AgentWorldBench - MCPtie54.955.3tie
Questions people ask
Which is better, Claude Opus 4.8 or Qwen3.6 Plus?
Claude Opus 4.8 wins six of the eight areas we test: coding, agents, reasoning, facts, images and charts and math. Qwen3.6 Plus wins none, but costs 88% less. They are level on long documents and following instructions.
Which is better for coding?
Claude Opus 4.8. It wins 9 of the 9 coding tests both models report; Qwen3.6 Plus wins none.
Which is cheaper?
Claude Opus 4.8 costs $5.00 per million input tokens and $25.00 per million output tokens; Qwen3.6 Plus costs $0.50 and $3.00. That makes Qwen3.6 Plus about 88% 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 29 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.