Claude Sonnet 4.6 vs GPT-5
Wins 5 of 7 areas
Coding · Agents · Reasoning · Facts · Long documents
Wins 1 of 7 areas
Following instructions
Claude Sonnet 4.6 is the stronger all-rounder.GPT-5 is cheaper and better at following instructions.
Scores updated · 36 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 software60Claude Sonnet 4.66 of 6 tests
- AgentsCarrying out multi-step tasks on its own30Claude Sonnet 4.63 of 3 tests
- ReasoningHard problems that need careful thinking31Claude Sonnet 4.63 of 4 tests
- FactsGetting facts right instead of making them up21Claude Sonnet 4.62 of 4 tests · 1 tie
- Long documentsFinding answers in very long texts10Claude Sonnet 4.61 of 1 test
- Following instructionsDoing exactly what it is asked01GPT-51 of 1 test
- Images and chartsUnderstanding pictures, charts and video11Even1 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.
GPT-5 costs 38% 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
- Abstract visual puzzles that people can solveARC-AGI-2+48.4points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+36points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+33.8points ahead
Where GPT-5 pulls ahead
- Short factual questions, answered correctlySimpleQA Verified+17.3points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+16.5points ahead
- Reasoning about charts from research papersCharXiv (RQ)+9.5points ahead
Every test, side by side
All 36 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Sonnet 4.6
- Terminal-Bench 2.1Claude Sonnet 4.6 by 3671.235.2+36
- Terminal-Bench HardClaude Sonnet 4.6 by 20.45332.6+20.4
- SWE-bench ProClaude Sonnet 4.6 by 16.358.141.8+16.3
- LMArena · WebDevClaude Sonnet 4.6 by 104 rating points15211417+104 rating
- SciCodeClaude Sonnet 4.6 by 7.250.142.9+7.2
- SWE-bench VerifiedClaude Sonnet 4.6 by 4.779.674.9+4.7
AgentsClaude Sonnet 4.6
- BrowseCompClaude Sonnet 4.6 by 19.874.754.9+19.8
- GDPValClaude Sonnet 4.6 by 15.736.721+15.7
- τ-Bench V3 · BankingClaude Sonnet 4.6 by 12.334.422.1+12.3
ReasoningClaude Sonnet 4.6
- ARC-AGI-2Claude Sonnet 4.6 by 48.458.39.9+48.4
- Humanity's Last ExamClaude Sonnet 4.6 by 5.133.628.5+5.1
- CritPtGPT-5 by 2.63.15.7+2.6
- GPQA DiamondClaude Sonnet 4.6 by 2.187.585.4+2.1
FactsClaude Sonnet 4.6
- AA-Omniscience · Non-hallucinationClaude Sonnet 4.6 by 33.851.617.8+33.8
- SimpleQA VerifiedGPT-5 by 17.332.850.1+17.3
- Vectara HHEM hallucination ratelower is betterClaude Sonnet 4.6 by 4.510.615.1+4.5
- AA-Omniscience · Accuracytie40.940.3tie
Images and chartsEven
- CharXiv (RQ)GPT-5 by 9.571.681.1+9.5
- LMArena · VisionClaude Sonnet 4.6 by 75 rating points12831208+75 rating
- MMMU-Protie73.374.2tie
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.
- Artificial Analysis Coding IndexClaude Sonnet 4.6 by 25.26337.8+25.2
- Terminal-Bench 2.0Claude Sonnet 4.6 by 23.959.135.2+23.9
- AA-OmniscienceClaude Sonnet 4.6 by 20.912.2-8.7+20.9
- ARC-AGI-1Claude Sonnet 4.6 by 20.38665.7+20.3
- HLE (with tools)Claude Sonnet 4.6 by 11.646.835.2+11.6
- τ²-Bench (Retail)Claude Sonnet 4.6 by 10.691.781.1+10.6
- τ²-Bench Telecom (AA run)GPT-5 by 9.175.784.8+9.1
- AA IntelligenceClaude Sonnet 4.6 by 7.130.123+7.1
- AA Agentic IndexClaude Sonnet 4.6 by 6.633.126.5+6.6
- vectara_avg_summary_lengthGPT-5 by 48 rating points114.7162.7+48 rating
- vectara_factual_consistencyClaude Sonnet 4.6 by 4.589.484.9+4.5
- frontiermath_tier_4_v1GPT-5 by 4.28.312.5+4.2
- GPQA (unspecified)Claude Sonnet 4.6 by 4.289.985.7+4.2
- vectara_answer_ratetie99.999.9tie
Questions people ask
Which is better, Claude Sonnet 4.6 or GPT-5?
Claude Sonnet 4.6 wins five of the seven areas where both have results: coding, agents, reasoning, facts and long documents. GPT-5 wins following instructions, and costs 38% less. They are level on images and charts.
Which is better for coding?
Claude Sonnet 4.6. It wins 6 of the 6 coding tests both models report; GPT-5 wins none.
Which is cheaper?
Claude Sonnet 4.6 costs $3.00 per million input tokens and $15.00 per million output tokens; GPT-5 costs $1.25 and $10.00. That makes GPT-5 about 38% cheaper for the same work.
How do you compare the two?
We use the 36 benchmark tests both models have published scores on. The verdict counts the 22 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.