GPT-5.6 Luna vs Claude Sonnet 5.5
Wins 2 of 8 areas
Long documents · Following instructions
Wins 6 of 8 areas
Coding · Agents · Reasoning · Facts · Images and charts · Math
Claude Sonnet 5.5 is the stronger all-rounder.GPT-5.6 Luna is cheaper and better at long documents.
Scores updated · 34 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.
- ReasoningHard problems that need careful thinking06Claude Sonnet 5.56 of 6 tests
- CodingWriting and fixing software05Claude Sonnet 5.55 of 5 tests
- FactsGetting facts right instead of making them up03Claude Sonnet 5.53 of 3 tests
- MathCompetition and research-level math03Claude Sonnet 5.53 of 3 tests
- AgentsCarrying out multi-step tasks on its own02Claude Sonnet 5.52 of 2 tests
- Images and chartsUnderstanding pictures, charts and video01Claude Sonnet 5.51 of 1 test
- Long documentsFinding answers in very long texts10GPT-5.6 Luna1 of 1 test
- Following instructionsDoing exactly what it is asked10GPT-5.6 Luna1 of 1 test
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.
GPT-5.6 Luna costs 88% 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 5.5 pulls ahead
- Complex command-line tasks across many fieldsTerminal-Bench 4.0+52points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+45.6points ahead
- Common-sense trick questionsSimpleBench+29.1points ahead
Where GPT-5.6 Luna pulls ahead
- Follows detailed instructions exactlyLiveBench · Instruction Following+3.3points ahead
- Reasons across sets of long documentsAA-LCR+1points ahead
Every test, side by side
All 34 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Sonnet 5.5
- LMArena · WebDevClaude Sonnet 5.5 by 251 rating points15191770+251 rating
- SWE-bench ProClaude Sonnet 5.5 by 18.662.781.3+18.6
- LiveBench · CodingClaude Sonnet 5.5 by 8.582.991.4+8.5
- LiveBench · Agentic CodingClaude Sonnet 5.5 by 7.948.456.3+7.9
- SciCodeClaude Sonnet 5.5 by 7.453.661+7.4
AgentsClaude Sonnet 5.5
- Terminal-Bench 4.0Claude Sonnet 5.5 by 5211.663.6+52
- GDPValClaude Sonnet 5.5 by 18.94866.9+18.9
ReasoningClaude Sonnet 5.5
- SimpleBenchClaude Sonnet 5.5 by 29.146.875.9+29.1
- ARC-AGI-2Claude Sonnet 5.5 by 24.359.683.9+24.3
- ARC-AGI-3Claude Sonnet 5.5 by 24.10.224.3+24.1
- Humanity's Last ExamClaude Sonnet 5.5 by 15.539.555+15.5
- CritPtClaude Sonnet 5.5 by 10.820.631.4+10.8
- LiveBench · ReasoningClaude Sonnet 5.5 by 685.691.6+6
FactsClaude Sonnet 5.5
- AA-Omniscience · Non-hallucinationClaude Sonnet 5.5 by 45.67.453+45.6
- AA-Omniscience · AccuracyClaude Sonnet 5.5 by 11.342.754+11.3
- SimpleQA VerifiedClaude Sonnet 5.5 by 5.54146.5+5.5
Images and chartsClaude Sonnet 5.5
- LMArena · VisionClaude Sonnet 5.5 by 30 rating points12571287+30 rating
MathClaude Sonnet 5.5
- FrontierMath Tier 4Claude Sonnet 5.5 by 2258.580.5+22
- LiveBench · MathematicsClaude Sonnet 5.5 by 8.987.296.1+8.9
- FrontierMath Tiers 1-3 (v2)Claude Sonnet 5.5 by 6.782.188.8+6.7
Following instructionsGPT-5.6 Luna
- LiveBench · Instruction FollowingGPT-5.6 Luna by 3.360.156.8+3.3
Other results12 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.
- Terminal-Bench 4.0Claude Sonnet 5.5 by 53.317.370.6+53.3
- AA-OmniscienceClaude Sonnet 5.5 by 42.6-10.332.3+42.6
- AA IntelligenceClaude Sonnet 5.5 by 18.737.356+18.7
- livebench_data_analysisGPT-5.6 Luna by 18.57859.5+18.5
- HLE (with tools)Claude Sonnet 5.5 by 15.648.964.5+15.6
- HealthBench ProfessionalClaude Sonnet 5.5 by 13.555.769.2+13.5
- HealthBench Professional length-adjustedClaude Sonnet 5.5 by 13.555.769.2+13.5
- Toolathlon VerifiedClaude Sonnet 5.5 by 9.967.977.8+9.9
- HealthBenchClaude Sonnet 5.5 by 9.655.865.4+9.6
- ARC-AGI-1Claude Sonnet 5.5 by 5.890.796.5+5.8
- livebench_languageClaude Sonnet 5.5 by 5.472.678+5.4
- DeepSWE 1.1Claude Sonnet 5.5 by 46771+4
Questions people ask
Which is better, GPT-5.6 Luna or Claude Sonnet 5.5?
Claude Sonnet 5.5 wins six of the eight areas we test: coding, agents, reasoning, facts, images and charts and math. GPT-5.6 Luna wins long documents and following instructions, and costs 88% less.
Which is better for coding?
Claude Sonnet 5.5. It wins 5 of the 5 coding tests both models report; GPT-5.6 Luna wins none.
Which is cheaper?
GPT-5.6 Luna costs $0.20 per million input tokens and $1.20 per million output tokens; Claude Sonnet 5.5 costs $2.00 and $10.00. That makes GPT-5.6 Luna about 88% cheaper for the same work.
How do you compare the two?
We use the 34 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 12 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.