Claude Sonnet 5 vs Qwen3.8 27B
Wins 4 of 8 areas
Coding · Agents · Reasoning · Math
Wins 1 of 8 areas
Following instructions
Claude Sonnet 5 wins more areas, narrowly.Qwen3.8 27B is cheaper and better at following instructions.
Scores updated · 40 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 software52Claude Sonnet 55 of 8 tests · 1 tie
- ReasoningHard problems that need careful thinking30Claude Sonnet 53 of 5 tests · 2 ties
- AgentsCarrying out multi-step tasks on its own32Claude Sonnet 53 of 5 tests
- MathCompetition and research-level math10Claude Sonnet 51 of 1 test
- Following instructionsDoing exactly what it is asked01Qwen3.8 27B1 of 1 test
- FactsGetting facts right instead of making them up11Even1 each
- Images and chartsUnderstanding pictures, charts and video11Even1 each · 1 tie
- Long documentsFinding answers in very long texts00Even0 each · 1 tie
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.
Qwen3.8 27B costs 71% 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 5 pulls ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+24.5points ahead
- Finds hard-to-locate facts by browsing the webBrowseComp+23.1points ahead
- Research-level physics problemsCritPt+11.5points ahead
Where Qwen3.8 27B pulls ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+10.7points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+9.1points ahead
- Follows detailed instructions exactlyLiveBench · Instruction Following+8.8points ahead
Every test, side by side
All 40 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Sonnet 5
- SciCodeClaude Sonnet 5 by 7.754.346.6+7.7
- SWE-bench VerifiedClaude Sonnet 5 by 7.485.277.8+7.4
- LMArena · WebDevQwen3.8 27B by 51 rating points15391590+51 rating
- LiveBench · CodingClaude Sonnet 5 by 580.775.7+5
- SWE-bench MultilingualClaude Sonnet 5 by 4.578.373.8+4.5
- LiveBench · Agentic CodingQwen3.8 27B by 259.461.4+2
- SWE-bench ProClaude Sonnet 5 by 1.563.261.7+1.5
- Terminal-Bench 2.1tie80.579.8tie
AgentsClaude Sonnet 5
- BrowseCompClaude Sonnet 5 by 23.184.761.6+23.1
- τ-Bench V3 · BankingQwen3.8 27B by 10.737.348+10.7
- Terminal-Bench 4.0Claude Sonnet 5 by 8.514.15.6+8.5
- OSWorld-VerifiedQwen3.8 27B by 3.181.284.3+3.1
- GDPValClaude Sonnet 5 by 2.248.246+2.2
ReasoningClaude Sonnet 5
- CritPtClaude Sonnet 5 by 11.516.95.4+11.5
- LiveBench · ReasoningClaude Sonnet 5 by 8.788.780+8.7
- Humanity's Last ExamClaude Sonnet 5 by 7.441.333.9+7.4
- GPQA Diamondtie91.190.5tie
- SimpleBenchtie60.660.2tie
FactsEven
- AA-Omniscience · AccuracyClaude Sonnet 5 by 24.54015.6+24.5
- AA-Omniscience · Non-hallucinationQwen3.8 27B by 9.160.669.7+9.1
Images and chartsEven
- CharXiv (RQ)Qwen3.8 27B by 1.988.390.2+1.9
- MMMU-ProClaude Sonnet 5 by 177.376.3+1
- LMArena · Visiontie12721271tie
Following instructionsQwen3.8 27B
- LiveBench · Instruction FollowingQwen3.8 27B by 8.863.972.7+8.8
Other results14 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.
- GDP (Surge AI)Claude Sonnet 5 by 65.281.616.4+65.2
- AA-OmniscienceClaude Sonnet 5 by 26.516.4-10+26.5
- DeepSWE 1.1Claude Sonnet 5 by 11.85442.2+11.8
- SWE-bench MultimodalQwen3.8 27B by 10.528.138.6+10.5
- Toolathlon VerifiedClaude Sonnet 5 by 7.674.767.1+7.6
- Terminal-Bench 4.0Claude Sonnet 5 by 6.812.45.6+6.8
- OSWorld 2.0 (partial)Qwen3.8 27B by 5.442.648+5.4
- livebench_data_analysisQwen3.8 27B by 4.971.776.6+4.9
- AA IntelligenceClaude Sonnet 5 by 4.538.233.7+4.5
- LVBenchQwen3.8 27B by 3.968.572.4+3.9
- Artificial Analysis Coding IndexClaude Sonnet 5 by 3.471.568.1+3.4
- AA Agentic IndexQwen3.8 27B by 2.244.346.5+2.2
- GDPval-AA 2.1Claude Sonnet 5 by 22 rating points14491427+22 rating
- livebench_languagetie7574.3tie
Questions people ask
Which is better, Claude Sonnet 5 or Qwen3.8 27B?
Claude Sonnet 5 wins four of the eight areas we test: coding, agents, reasoning and math. Qwen3.8 27B wins following instructions, and costs 71% less. They are level on facts, images and charts and long documents.
Which is better for coding?
Claude Sonnet 5. It wins 5 of the 8 coding tests both models report; Qwen3.8 27B wins 2, and 1 is a tie.
Which is cheaper?
Claude Sonnet 5 costs $2.00 per million input tokens and $10.00 per million output tokens; Qwen3.8 27B costs $0.50 and $3.00. That makes Qwen3.8 27B about 71% cheaper for the same work.
How do you compare the two?
We use the 40 benchmark tests both models have published scores on. The verdict counts the 26 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 14 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.