Claude Opus 4.8 vs Qwen3.7 Max
Wins 4 of 7 areas
Coding · Agents · Reasoning · Math
Wins 3 of 7 areas
Facts · Long documents · Following instructions
Claude Opus 4.8 wins more areas, narrowly.Qwen3.7 Max is cheaper and better at following instructions.
Scores updated · 54 tests both models report · How we compare
Where each one wins
Tests won in each of the seven areas where both have results. 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
- AgentsCarrying out multi-step tasks on its own40Claude Opus 4.84 of 4 tests
- MathCompetition and research-level math42Claude Opus 4.84 of 6 tests
- ReasoningHard problems that need careful thinking31Claude Opus 4.83 of 5 tests · 1 tie
- Following instructionsDoing exactly what it is asked02Qwen3.7 Max2 of 2 tests
- FactsGetting facts right instead of making them up12Qwen3.7 Max2 of 3 tests
- Long documentsFinding answers in very long texts01Qwen3.7 Max1 of 1 test
Long documents 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.
Qwen3.7 Max costs 67% 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
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+22.4points ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+21.9points ahead
- Complex command-line tasks across many fieldsTerminal-Bench 4.0+20.2points ahead
Where Qwen3.7 Max pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+18.3points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+13.7points ahead
- Olympiad math problems with short answersIMOAnswerBench+6.5points ahead
Every test, side by side
All 54 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 10.184.674.5+10.1
- SWE-bench ProClaude Opus 4.8 by 8.669.260.6+8.6
- SWE-bench VerifiedClaude Opus 4.8 by 8.288.680.4+8.2
- LiveBench · CodingClaude Opus 4.8 by 7.681.874.2+7.6
- Terminal-Bench HardClaude Opus 4.8 by 7.558.350.8+7.5
- LiveBench · Agentic CodingClaude Opus 4.8 by 6.950.543.6+6.9
- SWE-bench MultilingualClaude Opus 4.8 by 6.184.478.3+6.1
- SciCodeClaude Opus 4.8 by 4.954.449.5+4.9
- LMArena · WebDevClaude Opus 4.8 by 41 rating points15561515+41 rating
AgentsClaude Opus 4.8
- τ-Bench V3 · BankingClaude Opus 4.8 by 22.434.211.8+22.4
- Terminal-Bench 4.0Claude Opus 4.8 by 20.221.71.5+20.2
- GDPValClaude Opus 4.8 by 16.247.831.6+16.2
- MCP AtlasClaude Opus 4.8 by 5.882.276.4+5.8
ReasoningClaude Opus 4.8
- Humanity's Last ExamClaude Opus 4.8 by 8.248.740.5+8.2
- CritPtClaude Opus 4.8 by 7.520.913.4+7.5
- LiveBench · ReasoningClaude Opus 4.8 by 5.989.283.3+5.9
- SimpleBenchQwen3.7 Max by 5.664.870.4+5.6
- GPQA Diamondtie9292.3tie
FactsQwen3.7 Max
- AA-Omniscience · AccuracyClaude Opus 4.8 by 17.748.831.1+17.7
- AA-Omniscience · Non-hallucinationQwen3.7 Max by 13.760.774.4+13.7
- SimpleQA VerifiedQwen3.7 Max by 2.85355.8+2.8
MathClaude Opus 4.8
- FrontierMath Tier 4Claude Opus 4.8 by 21.956.134.1+21.9
- FrontierMath Tiers 1-3 (v2)Claude Opus 4.8 by 15.48064.6+15.4
- LiveBench · MathematicsClaude Opus 4.8 by 994.385.3+9
- IMOAnswerBenchQwen3.7 Max by 6.583.590+6.5
- AIME 2026Claude Opus 4.8 by 310097+3
- HMMT Feb. 2026Qwen3.7 Max by 1.695.597.1+1.6
Following instructionsQwen3.7 Max
- IFBenchQwen3.7 Max by 18.362.280.5+18.3
- LiveBench · Instruction FollowingQwen3.7 Max by 27274+2
Other results24 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.
- DeepSWEClaude Opus 4.8 by 405818+40
- DeepSWE 1.1Claude Opus 4.8 by 37.45921.6+37.4
- Toolathlon VerifiedClaude Opus 4.8 by 30.279.949.7+30.2
- NL2RepoClaude Opus 4.8 by 22.569.747.2+22.5
- AA Agentic IndexClaude Opus 4.8 by 18.742.623.9+18.7
- JobBenchClaude Opus 4.8 by 17.148.431.3+17.1
- PaperBenchClaude Opus 4.8 by 15.580.364.8+15.5
- AA-OmniscienceClaude Opus 4.8 by 15.328.813.5+15.3
- Agents' Last Exam (Pass / Score)Claude Opus 4.8 by 1445.131.1+14
- AndroidBenchClaude Opus 4.8 by 13.369.856.5+13.3
- AA IntelligenceClaude Opus 4.8 by 12.341.829.5+12.3
- MLS Bench Litelower is betterQwen3.7 Max by 11.142.831.7+11.1
- Artificial Analysis Coding IndexClaude Opus 4.8 by 8.374.366+8.3
- livebench_data_analysisQwen3.7 Max by 5.86671.8+5.8
- Finance Agent v2Claude Opus 4.8 by 5.553.948.4+5.5
- Terminal-Bench 2.0Claude Opus 4.8 by 4.974.669.7+4.9
- HealthBenchClaude Opus 4.8 by 4.859.354.5+4.8
- HLE (with tools)Claude Opus 4.8 by 4.457.953.5+4.4
- LiveBenchClaude Opus 4.8 by 2.977.274.3+2.9
- WideSearchQwen3.7 Max by 2.372.975.2+2.3
- HMMT Nov. 2025Claude Opus 4.8 by 1.596.595+1.5
- IncludeClaude Opus 4.8 by 1.487.686.2+1.4
- τ²-Bench Telecom (AA run)tie94.494.7tie
- livebench_languagetie79.779.7tie
Questions people ask
Which is better, Claude Opus 4.8 or Qwen3.7 Max?
Claude Opus 4.8 wins four of the seven areas where both have results: coding, agents, reasoning and math. Qwen3.7 Max wins facts, long documents and following instructions, and costs 67% less.
Which is better for coding?
Claude Opus 4.8. It wins 9 of the 9 coding tests both models report; Qwen3.7 Max wins none.
Which is cheaper?
Claude Opus 4.8 costs $5.00 per million input tokens and $25.00 per million output tokens; Qwen3.7 Max costs $2.50 and $7.50. That makes Qwen3.7 Max about 67% cheaper for the same work.
How do you compare the two?
We use the 54 benchmark tests both models have published scores on. The verdict counts the 30 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 24 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.