Claude Sonnet 5 vs GPT-5.6 Sol
Wins 0 of 8 areas
—
Wins 7 of 8 areas
Coding · Agents · Reasoning · Facts · Math · Long documents · Following instructions
GPT-5.6 Sol is the stronger all-rounder.Claude Sonnet 5 is cheaper.
Scores updated · 51 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 own05GPT-5.6 Sol5 of 5 tests
- ReasoningHard problems that need careful thinking05GPT-5.6 Sol5 of 5 tests
- CodingWriting and fixing software15GPT-5.6 Sol5 of 6 tests
- MathCompetition and research-level math03GPT-5.6 Sol3 of 3 tests
- FactsGetting facts right instead of making them up12GPT-5.6 Sol2 of 3 tests
- Long documentsFinding answers in very long texts01GPT-5.6 Sol1 of 1 test
- Following instructionsDoing exactly what it is asked01GPT-5.6 Sol1 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.
Claude Sonnet 5 costs 50% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where GPT-5.6 Sol pulls ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+53.7points ahead
- Short factual questions, answered correctlySimpleQA Verified+36points ahead
- Complex command-line tasks across many fieldsTerminal-Bench 4.0+25.8points ahead
Where Claude Sonnet 5 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+52.8points ahead
- Fixes real GitHub issues, working as an agentLiveBench · Agentic Coding+3.2points ahead
- Reasoning about charts from research papersCharXiv (RQ)+2.5points ahead
Every test, side by side
All 51 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGPT-5.6 Sol
- LMArena · WebDevGPT-5.6 Sol by 77 rating points15411618+77 rating
- Terminal-Bench 2.1GPT-5.6 Sol by 7.580.588+7.5
- LiveBench · Agentic CodingClaude Sonnet 5 by 3.259.456.2+3.2
- LiveBench · CodingGPT-5.6 Sol by 3.280.783.9+3.2
- SciCodeGPT-5.6 Sol by 2.854.357.1+2.8
- SWE-bench ProGPT-5.6 Sol by 1.463.264.6+1.4
AgentsGPT-5.6 Sol
- Terminal-Bench 4.0GPT-5.6 Sol by 25.814.139.9+25.8
- GDPValGPT-5.6 Sol by 7.348.355.6+7.3
- τ-Bench V3 · BankingGPT-5.6 Sol by 737.344.3+7
- BrowseCompGPT-5.6 Sol by 5.784.790.4+5.7
- OSWorld-VerifiedGPT-5.6 Sol by 1.881.283+1.8
ReasoningGPT-5.6 Sol
- CritPtGPT-5.6 Sol by 15.416.932.3+15.4
- Humanity's Last ExamGPT-5.6 Sol by 8.241.349.5+8.2
- SimpleBenchGPT-5.6 Sol by 4.260.664.8+4.2
- GPQA DiamondGPT-5.6 Sol by 391.194.1+3
- LiveBench · ReasoningGPT-5.6 Sol by 388.791.7+3
FactsGPT-5.6 Sol
- AA-Omniscience · Non-hallucinationClaude Sonnet 5 by 52.860.67.8+52.8
- SimpleQA VerifiedGPT-5.6 Sol by 3633.769.7+36
- AA-Omniscience · AccuracyGPT-5.6 Sol by 19.34059.4+19.3
Images and chartsEven
- MMMU-ProGPT-5.6 Sol by 6.177.383.4+6.1
- CharXiv (RQ)Claude Sonnet 5 by 2.588.385.8+2.5
- LMArena · Visiontie12751279tie
MathGPT-5.6 Sol
- FrontierMath Tier 4GPT-5.6 Sol by 53.729.383+53.7
- FrontierMath Tiers 1-3 (v2)GPT-5.6 Sol by 23.565.689.1+23.5
- LiveBench · MathematicsGPT-5.6 Sol by 3.392.996.2+3.3
Following instructionsGPT-5.6 Sol
- LiveBench · Instruction FollowingGPT-5.6 Sol by 863.971.8+8
Other results24 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.
- Program BenchClaude Sonnet 5 by 52.377.325+52.3
- GDP (Surge AI)Claude Sonnet 5 by 50.981.630.7+50.9
- Terminal-Bench 4.0GPT-5.6 Sol by 24.912.437.3+24.9
- HLE-VerifiedGPT-5.6 Sol by 23.53154.5+23.5
- OSWorld 2.0 (partial)GPT-5.6 Sol by 2042.662.6+20
- Terminal-bench 3.0GPT-5.6 Sol by 19.814.634.4+19.8
- DeepSWE 1.1GPT-5.6 Sol by 195473+19
- GDPval-AA 2.1GPT-5.6 Sol by 139 rating points14491588+139 rating
- LVBenchGPT-5.6 Sol by 13.668.582.1+13.6
- livebench_languageGPT-5.6 Sol by 12.77587.7+12.7
- GDPval-AA v2 EloGPT-5.6 Sol by 126 rating points15841710+126 rating
- GDP.PDF (All pass rate)GPT-5.6 Sol by 122840+12
- GDPval-AA v2GPT-5.6 Sol by 93 rating points16181711+93 rating
- AA IntelligenceGPT-5.6 Sol by 8.838.247+8.8
- livebench_data_analysisGPT-5.6 Sol by 8.171.779.8+8.1
- HLE (with tools)GPT-5.6 Sol by 7.157.464.5+7.1
- AA Agentic IndexGPT-5.6 Sol by 6.244.350.5+6.2
- Artificial Analysis Coding IndexGPT-5.6 Sol by 5.971.577.4+5.9
- AA-OmniscienceGPT-5.6 Sol by 5.516.422+5.5
- OfficeQA ProGPT-5.6 Sol by 3.859.463.2+3.8
- ToolathlonGPT-5.6 Sol by 3.754.358+3.7
- HealthBench ProfessionalGPT-5.6 Sol by 2.757.860.5+2.7
- ProteinGym HardClaude Sonnet 5 by 1.136.635.5+1.1
- Toolathlon Verifiedtie74.774.9tie
Questions people ask
Which is better, Claude Sonnet 5 or GPT-5.6 Sol?
GPT-5.6 Sol wins seven of the eight areas we test: coding, agents, reasoning, facts, math, long documents and following instructions. Claude Sonnet 5 wins none, but costs 50% less. They are level on images and charts.
Which is better for coding?
GPT-5.6 Sol. It wins 5 of the 6 coding tests both models report; Claude Sonnet 5 wins 1.
Which is cheaper?
Claude Sonnet 5 costs $2.00 per million input tokens and $10.00 per million output tokens; GPT-5.6 Sol costs $4.00 and $20.00. That makes Claude Sonnet 5 about 50% cheaper for the same work.
How do you compare the two?
We use the 51 benchmark tests both models have published scores on. The verdict counts the 27 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 24 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.