Claude Opus 4.6 vs GPT-5.5
Wins 0 of 8 areas
—
Wins 8 of 8 areas
Coding · Agents · Reasoning · Facts · Images and charts · Math · Long documents · Following instructions
GPT-5.5 is the stronger all-rounder.Claude Opus 4.6 is cheaper.
Scores updated · 65 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 thinking06GPT-5.56 of 7 tests · 1 tie
- MathCompetition and research-level math06GPT-5.56 of 6 tests
- CodingWriting and fixing software15GPT-5.55 of 6 tests
- AgentsCarrying out multi-step tasks on its own13GPT-5.53 of 5 tests · 1 tie
- FactsGetting facts right instead of making them up13GPT-5.53 of 4 tests
- Following instructionsDoing exactly what it is asked02GPT-5.52 of 2 tests
- Images and chartsUnderstanding pictures, charts and video12GPT-5.52 of 3 tests
- Long documentsFinding answers in very long texts01GPT-5.51 of 1 test
Long documents rests on a single test.
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 Opus 4.6 costs 14% 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-5.5 pulls ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+45.7points ahead
- Proof problems from the US Math OlympiadUSAMO 2026+32points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+22.8points ahead
Where Claude Opus 4.6 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+26.2points ahead
- Real work tasks from 44 professionsGDPVal+13.2points ahead
Every test, side by side
All 65 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGPT-5.5
- Terminal-Bench HardGPT-5.5 by 14.446.260.6+14.4
- SWE-bench ProGPT-5.5 by 5.253.458.6+5.2
- LiveBench · Agentic CodingGPT-5.5 by 54954+5
- LiveBench · CodingGPT-5.5 by 478.282.2+4
- SciCodeGPT-5.5 by 3.951.955.8+3.9
- LMArena · WebDevClaude Opus 4.6 by 33 rating points15461513+33 rating
AgentsGPT-5.5
- GDPValClaude Opus 4.6 by 13.255.942.7+13.2
- MCP AtlasGPT-5.5 by 12.662.775.3+12.6
- OSWorld-VerifiedGPT-5.5 by 672.778.7+6
- AA ApexAgentsGPT-5.5 by 4.73337.7+4.7
- BrowseComptie8484.4tie
ReasoningGPT-5.5
- ARC-AGI-2GPT-5.5 by 16.268.885+16.2
- CritPtGPT-5.5 by 14.512.627.1+14.5
- Humanity's Last ExamGPT-5.5 by 5.939.945.8+5.9
- GPQA DiamondGPT-5.5 by 3.989.693.5+3.9
- SimpleBenchGPT-5.5 by 1.467.669+1.4
- LiveBench · ReasoningGPT-5.5 by 188.789.7+1
- ARC-AGI-3tie0.50.4tie
FactsGPT-5.5
- AA-Omniscience · Non-hallucinationClaude Opus 4.6 by 26.237.211+26.2
- SimpleQA VerifiedGPT-5.5 by 164763+16
- AA-Omniscience · AccuracyGPT-5.5 by 114758+11
- Vectara HHEM hallucination ratelower is betterGPT-5.5 by 2.912.29.3+2.9
Images and chartsGPT-5.5
- OCRBenchv2GPT-5.5 by 12.748.461.1+12.7
- MMMU-ProGPT-5.5 by 4.575.479.9+4.5
- LMArena · VisionClaude Opus 4.6 by 24 rating points13161292+24 rating
MathGPT-5.5
- FrontierMath Tier 4GPT-5.5 by 45.726.872.5+45.7
- USAMO 2026GPT-5.5 by 3266.298.2+32
- FrontierMath Tiers 1-3 (v2)GPT-5.5 by 19.36685.3+19.3
- LiveBench · MathematicsGPT-5.5 by 6.689.395.9+6.6
- AIME 2026GPT-5.5 by 3.396.7100+3.3
- HMMT Feb. 2026GPT-5.5 by 2.396.298.5+2.3
Following instructionsGPT-5.5
- IFBenchGPT-5.5 by 22.853.175.9+22.8
- LiveBench · Instruction FollowingGPT-5.5 by 7.463.370.7+7.4
Other results31 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.
- BabyVisionGPT-5.5 by 41.114.855.9+41.1
- AA Agentic IndexClaude Opus 4.6 by 30.367.637.3+30.3
- Artificial Analysis Coding IndexGPT-5.5 by 26.848.174.9+26.8
- ERQAGPT-5.5 by 23.740.864.5+23.7
- matharena_visual_math_overallGPT-5.5 by 22.672.394.9+22.6
- Terminal-Bench 2.0GPT-5.5 by 17.365.482.7+17.3
- GDPval-AA (Elo)GPT-5.5 by 150 rating points16191769+150 rating
- frontiermath_tier_4_v1GPT-5.5 by 12.522.935.4+12.5
- livebench_data_analysisGPT-5.5 by 11.769.981.6+11.7
- Tool-DecathlonGPT-5.5 by 8.447.255.6+8.4
- RealWorldQAGPT-5.5 by 8.373.982.2+8.3
- CyberGymGPT-5.5 by 873.881.8+8
- AA-OmniscienceGPT-5.5 by 6.813.720.5+6.8
- AA IntelligenceGPT-5.5 by 6.531.938.4+6.5
- LiveBenchGPT-5.5 by 4.476.380.7+4.4
- livebench_languageGPT-5.5 by 4.183.387.4+4.1
- OfficeQA ProClaude Opus 4.6 by 357.154.1+3
- vectara_factual_consistencyGPT-5.5 by 2.987.890.7+2.9
- Legal Agent BenchmarkClaude Opus 4.6 by 2.14.22.1+2.1
- ARC-AGI-1GPT-5.5 by 29395+2
- τ²-Bench Telecom (AA run)GPT-5.5 by 1.892.193.9+1.8
- JobBenchGPT-5.5 by 1.736.638.3+1.7
- HiL-Bench (Tools-allowed)GPT-5.5 by 1.438.339.7+1.4
- ToolathlonClaude Opus 4.6 by 1.256.855.6+1.2
- HLE (with tools)Claude Opus 4.6 by 1.153.352.2+1.1
- NL2Repotie49.850.7tie
- vectara_avg_summary_lengthtie137.6129.6tie
- BrowseComp (single-agent)tie83.784.4tie
- Finance Agenttie60.760tie
- HMMT Nov. 2025tie96.396.5tie
- vectara_answer_ratetie99.8100tie
Questions people ask
Which is better, Claude Opus 4.6 or GPT-5.5?
GPT-5.5 wins all eight areas we test: coding, agents, reasoning, facts, images and charts, math, long documents and following instructions. Claude Opus 4.6 wins none, but costs 14% less.
Which is better for coding?
GPT-5.5. It wins 5 of the 6 coding tests both models report; Claude Opus 4.6 wins 1.
Which is cheaper?
Claude Opus 4.6 costs $5.00 per million input tokens and $25.00 per million output tokens; GPT-5.5 costs $5.00 and $30.00. That makes Claude Opus 4.6 about 14% cheaper for the same work.
How do you compare the two?
We use the 65 benchmark tests both models have published scores on. The verdict counts the 34 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 31 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.