Claude Opus 4.5 vs Claude Sonnet 4.6
Wins 3 of 8 areas
Facts · Math · Following instructions
Wins 4 of 8 areas
Coding · Reasoning · Images and charts · Long documents
Claude Sonnet 4.6 wins more areas, narrowly.Claude Opus 4.5 is better at facts.
Scores updated · 45 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 software14Claude Sonnet 4.64 of 7 tests · 2 ties
- ReasoningHard problems that need careful thinking13Claude Sonnet 4.63 of 5 tests · 1 tie
- Images and chartsUnderstanding pictures, charts and video01Claude Sonnet 4.61 of 2 tests · 1 tie
- Long documentsFinding answers in very long texts01Claude Sonnet 4.61 of 1 test
- FactsGetting facts right instead of making them up21Claude Opus 4.52 of 4 tests · 1 tie
- Following instructionsDoing exactly what it is asked10Claude Opus 4.51 of 2 tests · 1 tie
- MathCompetition and research-level math10Claude Opus 4.51 of 1 test
- AgentsCarrying out multi-step tasks on its own22Even2 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.
Claude Sonnet 4.6 costs 40% 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
- Finds hard-to-locate facts by browsing the webBrowseComp+37.7points ahead
- Abstract visual puzzles that people can solveARC-AGI-2+20.7points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+12.6points ahead
Where Claude Opus 4.5 pulls ahead
- Short factual questions, answered correctlySimpleQA Verified+12.9points ahead
- Real work tasks from 44 professionsGDPVal+10.6points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+5.7points ahead
Every test, side by side
All 45 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Sonnet 4.6
- SWE-bench ProClaude Sonnet 4.6 by 6.15258.1+6.1
- Terminal-Bench HardClaude Sonnet 4.6 by 64753+6
- LMArena · WebDevClaude Sonnet 4.6 by 53 rating points14691522+53 rating
- LiveBench · Agentic CodingClaude Sonnet 4.6 by 2.939.742.6+2.9
- SWE-bench VerifiedClaude Opus 4.5 by 1.380.979.6+1.3
- SciCodetie49.550.1tie
- LiveBench · Codingtie79.779.3tie
AgentsEven
- BrowseCompClaude Sonnet 4.6 by 37.73774.7+37.7
- GDPValClaude Opus 4.5 by 10.647.336.7+10.6
- OSWorld-VerifiedClaude Sonnet 4.6 by 6.266.372.5+6.2
- MCP AtlasClaude Opus 4.5 by 162.361.3+1
ReasoningClaude Sonnet 4.6
- ARC-AGI-2Claude Sonnet 4.6 by 20.737.658.3+20.7
- LiveBench · ReasoningClaude Sonnet 4.6 by 4.780.184.8+4.7
- Humanity's Last ExamClaude Sonnet 4.6 by 3.530.133.6+3.5
- CritPtClaude Opus 4.5 by 1.54.63.1+1.5
- GPQA Diamondtie86.687.5tie
FactsClaude Opus 4.5
- SimpleQA VerifiedClaude Opus 4.5 by 12.945.732.8+12.9
- AA-Omniscience · Non-hallucinationClaude Sonnet 4.6 by 12.63951.6+12.6
- AA-Omniscience · AccuracyClaude Opus 4.5 by 5.746.640.9+5.7
- Vectara HHEM hallucination ratelower is bettertie10.910.6tie
Images and chartsClaude Sonnet 4.6
- CharXiv (RQ)Claude Sonnet 4.6 by 4.467.271.6+4.4
- MMMU-Protie7473.3tie
Following instructionsClaude Opus 4.5
- IFBenchClaude Opus 4.5 by 1.45856.6+1.4
- LiveBench · Instruction Followingtie62.563.2tie
Other results19 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 Agentic IndexClaude Opus 4.5 by 26.559.633.1+26.5
- Artificial Analysis Coding IndexClaude Sonnet 4.6 by 15.247.863+15.2
- τ²-Bench Telecom (AA run)Claude Opus 4.5 by 13.889.575.7+13.8
- Finance AgentClaude Sonnet 4.6 by 8.155.263.3+8.1
- livebench_languageClaude Opus 4.5 by 5.281.376.1+5.2
- frontiermath_tier_4_v1Claude Sonnet 4.6 by 4.14.28.3+4.1
- ARC-AGI-1Claude Sonnet 4.6 by 48286+4
- HLE (with tools)Claude Sonnet 4.6 by 3.643.246.8+3.6
- livebench_data_analysisClaude Sonnet 4.6 by 3.674.478+3.6
- τ²-Bench (Retail)Claude Sonnet 4.6 by 2.888.991.7+2.8
- ToolathlonClaude Opus 4.5 by 2.543.541+2.5
- AA-OmniscienceClaude Opus 4.5 by 1.81412.2+1.8
- MMMLUClaude Opus 4.5 by 1.590.889.3+1.5
- vectara_answer_rateClaude Sonnet 4.6 by 1.298.799.9+1.2
- AA IntelligenceClaude Sonnet 4.6 by 129.130.1+1
- LiveBenchtie7675.5tie
- vectara_factual_consistencytie89.189.4tie
- Terminal-Bench 2.0tie59.359.1tie
- vectara_avg_summary_lengthtie114.5114.7tie
Questions people ask
Which is better, Claude Opus 4.5 or Claude Sonnet 4.6?
Claude Sonnet 4.6 wins four of the eight areas we test: coding, reasoning, images and charts and long documents. Claude Opus 4.5 wins facts, math and following instructions. They are level on agents.
Which is better for coding?
Claude Sonnet 4.6. It wins 4 of the 7 coding tests both models report; Claude Opus 4.5 wins 1, and 2 are ties.
Which is cheaper?
Claude Opus 4.5 costs $5.00 per million input tokens and $25.00 per million output tokens; Claude Sonnet 4.6 costs $3.00 and $15.00. That makes Claude Sonnet 4.6 about 40% cheaper for the same work.
How do you compare the two?
We use the 45 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 19 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.