Claude Sonnet 4.6 vs GPT-5.6 Sol
Wins 0 of 8 areas
—
Wins 7 of 8 areas
Coding · Agents · Reasoning · Images and charts · Math · Long documents · Following instructions
GPT-5.6 Sol is the stronger all-rounder.Claude Sonnet 4.6 is cheaper.
Scores updated · 49 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.
- AgentsCarrying out multi-step tasks on its own08GPT-5.6 Sol8 of 8 tests
- CodingWriting and fixing software07GPT-5.6 Sol7 of 7 tests
- ReasoningHard problems that need careful thinking05GPT-5.6 Sol5 of 5 tests
- Images and chartsUnderstanding pictures, charts and video02GPT-5.6 Sol2 of 3 tests · 1 tie
- Following instructionsDoing exactly what it is asked02GPT-5.6 Sol2 of 2 tests
- MathCompetition and research-level math01GPT-5.6 Sol1 of 1 test
- Long documentsFinding answers in very long texts01GPT-5.6 Sol1 of 1 test
- FactsGetting facts right instead of making them up22Even2 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 25% 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 GPT-5.6 Sol pulls ahead
- Complex command-line tasks across many fieldsTerminal-Bench 4.0+36.9points ahead
- Short factual questions, answered correctlySimpleQA Verified+36.9points ahead
- Abstract visual puzzles that people can solveARC-AGI-2+34.2points ahead
Where Claude Sonnet 4.6 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+43.8points ahead
Every test, side by side
All 49 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGPT-5.6 Sol
- Terminal-Bench 2.1GPT-5.6 Sol by 16.871.288+16.8
- LiveBench · Agentic CodingGPT-5.6 Sol by 13.642.656.2+13.6
- Terminal-Bench HardGPT-5.6 Sol by 12.95365.9+12.9
- LMArena · WebDevGPT-5.6 Sol by 96 rating points15221618+96 rating
- SciCodeGPT-5.6 Sol by 750.157.1+7
- SWE-bench ProGPT-5.6 Sol by 6.558.164.6+6.5
- LiveBench · CodingGPT-5.6 Sol by 4.679.383.9+4.6
AgentsGPT-5.6 Sol
- Terminal-Bench 4.0GPT-5.6 Sol by 36.9339.9+36.9
- MCP AtlasGPT-5.6 Sol by 22.361.383.6+22.3
- GDPValGPT-5.6 Sol by 18.936.755.6+18.9
- AA IT-Bench SREGPT-5.6 Sol by 16.439.856.2+16.4
- BrowseCompGPT-5.6 Sol by 15.774.790.4+15.7
- AA ApexAgentsGPT-5.6 Sol by 11.92839.9+11.9
- OSWorld-VerifiedGPT-5.6 Sol by 10.572.583+10.5
- τ-Bench V3 · BankingGPT-5.6 Sol by 9.934.444.3+9.9
ReasoningGPT-5.6 Sol
- ARC-AGI-2GPT-5.6 Sol by 34.258.392.5+34.2
- CritPtGPT-5.6 Sol by 29.23.132.3+29.2
- Humanity's Last ExamGPT-5.6 Sol by 15.933.649.5+15.9
- LiveBench · ReasoningGPT-5.6 Sol by 6.984.891.7+6.9
- GPQA DiamondGPT-5.6 Sol by 6.687.594.1+6.6
FactsEven
- AA-Omniscience · Non-hallucinationClaude Sonnet 4.6 by 43.851.67.8+43.8
- SimpleQA VerifiedGPT-5.6 Sol by 36.932.869.7+36.9
- AA-Omniscience · AccuracyGPT-5.6 Sol by 18.540.959.4+18.5
- Vectara HHEM hallucination ratelower is betterClaude Sonnet 4.6 by 1.810.612.4+1.8
Images and chartsGPT-5.6 Sol
- CharXiv (RQ)GPT-5.6 Sol by 14.271.685.8+14.2
- MMMU-ProGPT-5.6 Sol by 10.173.383.4+10.1
- LMArena · Visiontie12831279tie
Following instructionsGPT-5.6 Sol
- IFBenchGPT-5.6 Sol by 16.156.672.7+16.1
- LiveBench · Instruction FollowingGPT-5.6 Sol by 8.763.271.8+8.7
Other results18 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.
- GDP (Surge AI)Claude Sonnet 4.6 by 47.978.630.7+47.9
- DeepSWE 1.1GPT-5.6 Sol by 433073+43
- GDPval-AA v2GPT-5.6 Sol by 316 rating points13951711+316 rating
- HLE (with tools)GPT-5.6 Sol by 17.746.864.5+17.7
- AA Agentic IndexGPT-5.6 Sol by 17.433.150.5+17.4
- ToolathlonGPT-5.6 Sol by 174158+17
- AA IntelligenceGPT-5.6 Sol by 16.930.147+16.9
- HealthBench ProfessionalGPT-5.6 Sol by 16.344.260.5+16.3
- Artificial Analysis Coding IndexGPT-5.6 Sol by 14.46377.4+14.4
- livebench_languageGPT-5.6 Sol by 11.676.187.7+11.6
- ARC-AGI-1GPT-5.6 Sol by 10.58696.5+10.5
- AA-OmniscienceGPT-5.6 Sol by 9.812.222+9.8
- OfficeQA ProGPT-5.6 Sol by 9.853.463.2+9.8
- τ²-Bench Telecom (AA run)GPT-5.6 Sol by 9.475.785.1+9.4
- livebench_data_analysisGPT-5.6 Sol by 1.87879.8+1.8
- vectara_factual_consistencyClaude Sonnet 4.6 by 1.889.487.6+1.8
- vectara_avg_summary_lengthGPT-5.6 Sol by 16.4 rating points114.7131.1+16.4 rating
- vectara_answer_ratetie99.999.1tie
Questions people ask
Which is better, Claude Sonnet 4.6 or GPT-5.6 Sol?
GPT-5.6 Sol wins seven of the eight areas we test: coding, agents, reasoning, images and charts, math, long documents and following instructions. Claude Sonnet 4.6 wins none, but costs 25% less. They are level on facts.
Which is better for coding?
GPT-5.6 Sol. It wins 7 of the 7 coding tests both models report; Claude Sonnet 4.6 wins none.
Which is cheaper?
Claude Sonnet 4.6 costs $3.00 per million input tokens and $15.00 per million output tokens; GPT-5.6 Sol costs $4.00 and $20.00. That makes Claude Sonnet 4.6 about 25% cheaper for the same work.
How do you compare the two?
We use the 49 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 18 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.