Claude Opus 4.8 vs Qwen3.8 27B
Wins 5 of 8 areas
Coding · Agents · Reasoning · Images and charts · Math
Wins 2 of 8 areas
Long documents · Following instructions
Claude Opus 4.8 is the stronger all-rounder.Qwen3.8 27B is cheaper and better at following instructions.
Scores updated · 38 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.
- ReasoningHard problems that need careful thinking60Claude Opus 4.86 of 6 tests
- CodingWriting and fixing software52Claude Opus 4.85 of 7 tests
- AgentsCarrying out multi-step tasks on its own21Claude Opus 4.82 of 4 tests · 1 tie
- Images and chartsUnderstanding pictures, charts and video10Claude Opus 4.81 of 2 tests · 1 tie
- MathCompetition and research-level math10Claude Opus 4.81 of 1 test
- Following instructionsDoing exactly what it is asked01Qwen3.8 27B1 of 2 tests · 1 tie
- Long documentsFinding answers in very long texts01Qwen3.8 27B1 of 1 test
- FactsGetting facts right instead of making them up11Even1 each
Math 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.
Qwen3.8 27B costs 88% 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 Opus 4.8 pulls ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+33.2points ahead
- Abstract visual puzzles that people can solveARC-AGI-2+29.7points ahead
- Complex command-line tasks across many fieldsTerminal-Bench 4.0+16.1points ahead
Where Qwen3.8 27B pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+17.3points ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+13.8points ahead
- Fixes real GitHub issues, working as an agentLiveBench · Agentic Coding+10.9points ahead
Every test, side by side
All 38 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Opus 4.8
- LiveBench · Agentic CodingQwen3.8 27B by 10.950.561.4+10.9
- SWE-bench MultilingualClaude Opus 4.8 by 10.684.473.8+10.6
- SciCodeClaude Opus 4.8 by 7.854.446.6+7.8
- SWE-bench ProClaude Opus 4.8 by 7.569.261.7+7.5
- LiveBench · CodingClaude Opus 4.8 by 6.181.875.7+6.1
- Terminal-Bench 2.1Claude Opus 4.8 by 4.884.679.8+4.8
- LMArena · WebDevQwen3.8 27B by 36 rating points15561592+36 rating
AgentsClaude Opus 4.8
- Terminal-Bench 4.0Claude Opus 4.8 by 16.121.75.6+16.1
- τ-Bench V3 · BankingQwen3.8 27B by 13.834.248+13.8
- GDPValClaude Opus 4.8 by 1.647.846.2+1.6
- OSWorld-Verifiedtie83.484.3tie
ReasoningClaude Opus 4.8
- ARC-AGI-2Claude Opus 4.8 by 29.772.142.4+29.7
- CritPtClaude Opus 4.8 by 15.520.95.4+15.5
- Humanity's Last ExamClaude Opus 4.8 by 14.848.733.9+14.8
- LiveBench · ReasoningClaude Opus 4.8 by 9.289.280+9.2
- SimpleBenchClaude Opus 4.8 by 4.664.860.2+4.6
- GPQA DiamondClaude Opus 4.8 by 1.59290.5+1.5
FactsEven
- AA-Omniscience · AccuracyClaude Opus 4.8 by 33.248.815.6+33.2
- AA-Omniscience · Non-hallucinationQwen3.8 27B by 960.769.7+9
Images and chartsClaude Opus 4.8
- LMArena · VisionClaude Opus 4.8 by 25 rating points12951270+25 rating
- CharXiv (RQ)tie89.990.2tie
Following instructionsQwen3.8 27B
- IFBenchQwen3.8 27B by 17.362.279.5+17.3
- LiveBench · Instruction Followingtie7272.7tie
Other results13 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-OmniscienceClaude Opus 4.8 by 38.828.8-10+38.8
- NL2RepoClaude Opus 4.8 by 27.469.742.3+27.4
- Agents' Last ExamQwen3.8 27B by 17.225.742.9+17.2
- DeepSWE 1.1Claude Opus 4.8 by 16.85942.2+16.8
- JobBenchClaude Opus 4.8 by 1548.433.4+15
- Toolathlon VerifiedClaude Opus 4.8 by 12.879.967.1+12.8
- livebench_data_analysisQwen3.8 27B by 10.66676.6+10.6
- AA IntelligenceClaude Opus 4.8 by 8.141.833.7+8.1
- Artificial Analysis Coding IndexClaude Opus 4.8 by 6.274.368.1+6.2
- livebench_languageClaude Opus 4.8 by 5.379.774.3+5.3
- ARC-AGI-1Claude Opus 4.8 by 592.587.5+5
- AA Agentic IndexQwen3.8 27B by 3.942.646.5+3.9
- SWE-bench Multimodaltie38.438.6tie
Questions people ask
Which is better, Claude Opus 4.8 or Qwen3.8 27B?
Claude Opus 4.8 wins five of the eight areas we test: coding, agents, reasoning, images and charts and math. Qwen3.8 27B wins long documents and following instructions, and costs 88% less. They are level on facts.
Which is better for coding?
Claude Opus 4.8. It wins 5 of the 7 coding tests both models report; Qwen3.8 27B wins 2.
Which is cheaper?
Claude Opus 4.8 costs $5.00 per million input tokens and $25.00 per million output tokens; Qwen3.8 27B costs $0.50 and $3.00. That makes Qwen3.8 27B about 88% cheaper for the same work.
How do you compare the two?
We use the 38 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 13 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.