Claude Opus 4.7 vs Claude Sonnet 5
Wins 5 of 8 areas
Coding · Facts · Images and charts · Math · Following instructions
Wins 2 of 8 areas
Agents · Long documents
Claude Opus 4.7 is the stronger all-rounder.Claude Sonnet 5 is cheaper and better at agents.
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 software61Claude Opus 4.76 of 8 tests · 1 tie
- Images and chartsUnderstanding pictures, charts and video30Claude Opus 4.73 of 3 tests
- MathCompetition and research-level math21Claude Opus 4.72 of 4 tests · 1 tie
- FactsGetting facts right instead of making them up21Claude Opus 4.72 of 3 tests
- Following instructionsDoing exactly what it is asked10Claude Opus 4.71 of 1 test
- AgentsCarrying out multi-step tasks on its own04Claude Sonnet 54 of 4 tests
- Long documentsFinding answers in very long texts02Claude Sonnet 52 of 2 tests
- ReasoningHard problems that need careful thinking22Even2 each · 1 tie
Following instructions 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.
Claude Sonnet 5 costs 60% 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 4.7 pulls ahead
- Short factual questions, answered correctlySimpleQA Verified+18points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+8.8points ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+4.6points ahead
Where Claude Sonnet 5 pulls ahead
- Finds one of eight look-alike replies in a long chatMRCR v2 8 needle 128k (average)+22.2points ahead
- Proof problems from the US Math OlympiadUSAMO 2026+10.2points ahead
- Fixes real GitHub issues, working as an agentLiveBench · Agentic Coding+8.7points 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 Opus 4.7
- LiveBench · Agentic CodingClaude Sonnet 5 by 8.750.759.4+8.7
- Terminal-Bench 2.1Claude Opus 4.7 by 2.683.180.5+2.6
- SWE-bench VerifiedClaude Opus 4.7 by 2.487.685.2+2.4
- SWE-bench MultilingualClaude Opus 4.7 by 2.280.578.3+2.2
- LMArena · WebDevClaude Opus 4.7 by 17 rating points15581541+17 rating
- LiveBench · CodingClaude Opus 4.7 by 1.482.180.7+1.4
- SWE-bench ProClaude Opus 4.7 by 1.164.363.2+1.1
- SciCodetie54.554.3tie
AgentsClaude Sonnet 5
- GDPValClaude Sonnet 5 by 5.542.848.3+5.5
- BrowseCompClaude Sonnet 5 by 5.479.384.7+5.4
- OSWorld-VerifiedClaude Sonnet 5 by 3.27881.2+3.2
- τ-Bench V3 · BankingClaude Sonnet 5 by 2.734.637.3+2.7
ReasoningEven
- CritPtClaude Sonnet 5 by 4.91216.9+4.9
- LiveBench · ReasoningClaude Sonnet 5 by 1.587.288.7+1.5
- SimpleBenchClaude Opus 4.7 by 1.161.760.6+1.1
- Humanity's Last ExamClaude Opus 4.7 by 142.341.3+1
- GPQA Diamondtie91.491.1tie
FactsClaude Opus 4.7
- SimpleQA VerifiedClaude Opus 4.7 by 1851.733.7+18
- AA-Omniscience · AccuracyClaude Opus 4.7 by 8.848.940+8.8
- AA-Omniscience · Non-hallucinationClaude Sonnet 5 by 2.957.760.6+2.9
Images and chartsClaude Opus 4.7
- LMArena · VisionClaude Opus 4.7 by 38 rating points13131275+38 rating
- CharXiv (RQ)Claude Opus 4.7 by 2.79188.3+2.7
- MMMU-ProClaude Opus 4.7 by 1.578.877.3+1.5
MathClaude Opus 4.7
- USAMO 2026Claude Sonnet 5 by 10.269.379.5+10.2
- FrontierMath Tiers 1-3 (v2)Claude Opus 4.7 by 4.670.265.6+4.6
- FrontierMath Tier 4Claude Opus 4.7 by 2.431.729.3+2.4
- LiveBench · Mathematicstie92.892.9tie
Long documentsClaude Sonnet 5
- MRCR v2 8 needle 128k (average)Claude Sonnet 5 by 22.259.381.5+22.2
- AA-LCRClaude Sonnet 5 by 3.378.782+3.3
Following instructionsClaude Opus 4.7
- LiveBench · Instruction FollowingClaude Opus 4.7 by 2.866.763.9+2.8
Other results15 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.
- OfficeQA ProClaude Opus 4.7 by 21.280.659.4+21.2
- Toolathlon Pass@∞Claude Opus 4.7 by 12.152.840.7+12.1
- OfficeQAClaude Opus 4.7 by 11.286.375.1+11.2
- Terminal-Bench 2.0Claude Sonnet 5 by 1169.480.4+11
- AA-OmniscienceClaude Opus 4.7 by 10.827.316.4+10.8
- livebench_data_analysisClaude Opus 4.7 by 6.678.371.7+6.6
- SWE-bench MultimodalClaude Opus 4.7 by 6.434.528.1+6.4
- ToolathlonClaude Opus 4.7 by 559.354.3+5
- AA Agentic IndexClaude Sonnet 5 by 4.839.544.3+4.8
- livebench_languageClaude Opus 4.7 by 2.977.975+2.9
- CharXiv Reasoning (With tools)Claude Opus 4.7 by 2.79188.3+2.7
- HLE (with tools)Claude Sonnet 5 by 2.754.757.4+2.7
- AA IntelligenceClaude Opus 4.7 by 2.540.738.2+2.5
- Artificial Analysis Coding IndexClaude Opus 4.7 by 2.173.671.5+2.1
- Legal Agent BenchmarkClaude Opus 4.7 by 1.37.15.8+1.3
Questions people ask
Which is better, Claude Opus 4.7 or Claude Sonnet 5?
Claude Opus 4.7 wins five of the eight areas we test: coding, facts, images and charts, math and following instructions. Claude Sonnet 5 wins agents and long documents, and costs 60% less. They are level on reasoning.
Which is better for coding?
Claude Opus 4.7. It wins 6 of the 8 coding tests both models report; Claude Sonnet 5 wins 1, and 1 is a tie.
Which is cheaper?
Claude Opus 4.7 costs $5.00 per million input tokens and $25.00 per million output tokens; Claude Sonnet 5 costs $2.00 and $10.00. That makes Claude Sonnet 5 about 60% 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 30 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 15 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.