Claude Opus 4 vs Claude Sonnet 4.6
Wins 0 of 6 areas
—
Wins 6 of 6 areas
Coding · Reasoning · Facts · Images and charts · Long documents · Following instructions
Claude Sonnet 4.6 is the stronger all-rounder.
Scores updated · 21 tests both models report · How we compare
Where each one wins
Tests won in each of the six 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 software03Claude Sonnet 4.63 of 3 tests
- ReasoningHard problems that need careful thinking03Claude Sonnet 4.63 of 3 tests
- FactsGetting facts right instead of making them up01Claude Sonnet 4.61 of 1 test
- Images and chartsUnderstanding pictures, charts and video01Claude Sonnet 4.61 of 1 test
- Long documentsFinding answers in very long texts01Claude Sonnet 4.61 of 1 test
- Following instructionsDoing exactly what it is asked01Claude Sonnet 4.61 of 1 test
Facts, images and charts, 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.
Claude Sonnet 4.6 costs 80% less for the same work.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where Claude Sonnet 4.6 pulls ahead
- Abstract visual puzzles that people can solveARC-AGI-2+49.7points ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+21.9points ahead
- Very hard expert questions across many subjectsHumanity's Last Exam+21.3points ahead
Where Claude Opus 4 pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 21 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Sonnet 4.6
- Terminal-Bench HardClaude Sonnet 4.6 by 21.931.153+21.9
- SciCodeClaude Sonnet 4.6 by 10.339.850.1+10.3
- SWE-bench VerifiedClaude Sonnet 4.6 by 7.172.579.6+7.1
ReasoningClaude Sonnet 4.6
- ARC-AGI-2Claude Sonnet 4.6 by 49.78.658.3+49.7
- Humanity's Last ExamClaude Sonnet 4.6 by 21.312.333.6+21.3
- GPQA DiamondClaude Sonnet 4.6 by 7.979.687.5+7.9
FactsClaude Sonnet 4.6
- Vectara HHEM hallucination ratelower is betterClaude Sonnet 4.6 by 1.41210.6+1.4
Images and chartsClaude Sonnet 4.6
- LMArena · VisionClaude Sonnet 4.6 by 91 rating points11921283+91 rating
Long documentsClaude Sonnet 4.6
- AA-LCRClaude Sonnet 4.6 by 10.769.380+10.7
Following instructionsClaude Sonnet 4.6
- IFBenchClaude Sonnet 4.6 by 2.953.756.6+2.9
Other results11 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.
- ARC-AGI-1Claude Sonnet 4.6 by 50.335.786+50.3
- Artificial Analysis Coding IndexClaude Sonnet 4.6 by 293463+29
- τ²-Bench (Retail)Claude Sonnet 4.6 by 9.981.891.7+9.9
- AA IntelligenceClaude Sonnet 4.6 by 9.520.630.1+9.5
- vectara_answer_rateClaude Sonnet 4.6 by 8.99199.9+8.9
- frontiermath_tier_4_v1Claude Sonnet 4.6 by 4.14.28.3+4.1
- τ²-Bench Telecom (AA run)Claude Sonnet 4.6 by 2.373.475.7+2.3
- vectara_factual_consistencyClaude Sonnet 4.6 by 1.48889.4+1.4
- LiveBenchtie74.675.5tie
- vectara_avg_summary_lengthtie123.2114.7tie
- MMMLUtie88.889.3tie
Questions people ask
Which is better, Claude Opus 4 or Claude Sonnet 4.6?
Claude Sonnet 4.6 wins all six areas where both have results: coding, reasoning, facts, images and charts, long documents and following instructions. Claude Opus 4 wins none.
Which is better for coding?
Claude Sonnet 4.6. It wins 3 of the 3 coding tests both models report; Claude Opus 4 wins none.
Which is cheaper?
Claude Opus 4 costs $15.00 per million input tokens and $75.00 per million output tokens; Claude Sonnet 4.6 costs $3.00 and $15.00. That makes Claude Sonnet 4.6 about 80% cheaper for the same work.
How do you compare the two?
We use the 21 benchmark tests both models have published scores on. The verdict counts the 10 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 11 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.