Claude Opus 4.7 vs Claude Opus 5
Wins 1 of 8 areas
Following instructions
Wins 5 of 8 areas
Coding · Agents · Reasoning · Facts · Math
Claude Opus 5 is the stronger all-rounder.Claude Opus 4.7 is better at following instructions.
Both rank among the ten best models we track in coding.
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 software07Claude Opus 57 of 8 tests · 1 tie
- ReasoningHard problems that need careful thinking07Claude Opus 57 of 7 tests
- AgentsCarrying out multi-step tasks on its own05Claude Opus 55 of 5 tests
- MathCompetition and research-level math03Claude Opus 53 of 3 tests
- FactsGetting facts right instead of making them up12Claude Opus 52 of 3 tests
- Following instructionsDoing exactly what it is asked10Claude Opus 4.71 of 1 test
- Images and chartsUnderstanding pictures, charts and video11Even1 each · 1 tie
- Long documentsFinding answers in very long texts00Even0 each · 1 tie
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.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Claude Opus 5 pulls ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+41.5points ahead
- Solves new puzzle games it has never seenARC-AGI-3+30points ahead
- Common-sense trick questionsSimpleBench+18.9points ahead
Where Claude Opus 4.7 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+18.5points ahead
- Reasoning about charts from research papersCharXiv (RQ)+7.3points ahead
- Follows detailed instructions exactlyLiveBench · Instruction Following+2.9points 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 5
- SWE-bench ProClaude Opus 5 by 14.964.379.2+14.9
- LiveBench · Agentic CodingClaude Opus 5 by 14.550.765.2+14.5
- LMArena · WebDevClaude Opus 5 by 134 rating points15581692+134 rating
- SWE-bench MultilingualClaude Opus 5 by 980.589.5+9
- SWE-bench VerifiedClaude Opus 5 by 8.487.696+8.4
- Terminal-Bench 2.1Claude Opus 5 by 683.189.1+6
- SciCodeClaude Opus 5 by 1.954.556.4+1.9
- LiveBench · Codingtie82.181.5tie
AgentsClaude Opus 5
- GDPValClaude Opus 5 by 18.442.861.2+18.4
- BrowseCompClaude Opus 5 by 11.579.390.8+11.5
- MCP AtlasClaude Opus 5 by 8.577.385.8+8.5
- τ-Bench V3 · BankingClaude Opus 5 by 7.534.642.1+7.5
- OSWorld-VerifiedClaude Opus 5 by 5.47883.4+5.4
ReasoningClaude Opus 5
- ARC-AGI-3Claude Opus 5 by 300.230.2+30
- SimpleBenchClaude Opus 5 by 18.961.780.6+18.9
- CritPtClaude Opus 5 by 17.11229.1+17.1
- ARC-AGI-2Claude Opus 5 by 14.675.890.4+14.6
- Humanity's Last ExamClaude Opus 5 by 12.642.354.9+12.6
- LiveBench · ReasoningClaude Opus 5 by 487.291.2+4
- GPQA DiamondClaude Opus 5 by 1.891.493.2+1.8
FactsClaude Opus 5
- AA-Omniscience · Non-hallucinationClaude Opus 4.7 by 18.557.739.2+18.5
- AA-Omniscience · AccuracyClaude Opus 5 by 1248.960.9+12
- SimpleQA VerifiedClaude Opus 5 by 8.251.759.9+8.2
Images and chartsEven
- CharXiv (RQ)Claude Opus 4.7 by 7.39183.7+7.3
- MMMU-ProClaude Opus 5 by 5.978.884.7+5.9
- LMArena · Visiontie13131321tie
MathClaude Opus 5
- FrontierMath Tier 4Claude Opus 5 by 41.531.773.2+41.5
- FrontierMath Tiers 1-3 (v2)Claude Opus 5 by 15.470.285.6+15.4
- LiveBench · MathematicsClaude Opus 5 by 2.892.895.7+2.8
Following instructionsClaude Opus 4.7
- LiveBench · Instruction FollowingClaude Opus 4.7 by 2.966.763.8+2.9
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.
- SWE-bench MultimodalClaude Opus 5 by 24.934.559.4+24.9
- AA Agentic IndexClaude Opus 5 by 16.739.556.2+16.7
- HiL-Bench (Tools-allowed)Claude Opus 5 by 15.341.757+15.3
- OfficeQA ProClaude Opus 4.7 by 13.780.666.9+13.7
- livebench_languageClaude Opus 5 by 10.877.988.7+10.8
- AA IntelligenceClaude Opus 5 by 10.140.750.8+10.1
- AA-OmniscienceClaude Opus 5 by 9.827.337.1+9.8
- HLE (with tools)Claude Opus 5 by 8.954.763.6+8.9
- OfficeQAClaude Opus 4.7 by 8.286.378.1+8.2
- ARC-AGI-1Claude Opus 5 by 5.59297.5+5.5
- Legal Agent BenchmarkClaude Opus 5 by 4.67.111.7+4.6
- GDPval-AA (Elo)Claude Opus 4.7 by 45 rating points17531708+45 rating
- Artificial Analysis Coding IndexClaude Opus 5 by 4.473.678+4.4
- livebench_data_analysisClaude Opus 4.7 by 3.778.374.5+3.7
Questions people ask
Which is better, Claude Opus 4.7 or Claude Opus 5?
Claude Opus 5 wins five of the eight areas we test: coding, agents, reasoning, facts and math. Claude Opus 4.7 wins following instructions. They are level on images and charts and long documents.
Which is better for coding?
Claude Opus 5. It wins 7 of the 8 coding tests both models report; Claude Opus 4.7 wins none, 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 Opus 5 costs $5.00 and $25.00. For a million tokens read plus a million written, they cost about the same.
How do you compare the two?
We use the 45 benchmark tests both models have published scores on. The verdict counts the 31 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.