GPT-5.2 vs Claude Opus 5
Wins 1 of 8 areas
Long documents
Wins 7 of 8 areas
Coding · Agents · Reasoning · Facts · Images and charts · Math · Following instructions
Claude Opus 5 is the stronger all-rounder.GPT-5.2 is cheaper and better at long documents.
Scores updated · 39 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.
- CodingWriting and fixing software07Claude Opus 57 of 7 tests
- ReasoningHard problems that need careful thinking06Claude Opus 56 of 6 tests
- AgentsCarrying out multi-step tasks on its own04Claude Opus 54 of 4 tests
- FactsGetting facts right instead of making them up03Claude Opus 53 of 3 tests
- Images and chartsUnderstanding pictures, charts and video03Claude Opus 53 of 3 tests
- MathCompetition and research-level math03Claude Opus 53 of 3 tests
- Following instructionsDoing exactly what it is asked01Claude Opus 51 of 1 test
- Long documentsFinding answers in very long texts10GPT-5.21 of 1 test
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-5.2 costs 48% 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 Opus 5 pulls ahead
- Long, multi-file coding tasks in real codebasesSWE-bench Pro+49.3points ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+41.5points ahead
- Abstract visual puzzles that people can solveARC-AGI-2+37.5points ahead
Where GPT-5.2 pulls ahead
- Reasons across sets of long documentsAA-LCR+3.4points ahead
Every test, side by side
All 39 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Opus 5
- SWE-bench ProClaude Opus 5 by 49.329.979.2+49.3
- LMArena · WebDevClaude Opus 5 by 276 rating points14161692+276 rating
- SWE-bench MultilingualClaude Opus 5 by 22.866.789.5+22.8
- SWE-bench VerifiedClaude Opus 5 by 168096+16
- LiveBench · Agentic CodingClaude Opus 5 by 14.950.365.2+14.9
- LiveBench · CodingClaude Opus 5 by 5.476.181.5+5.4
- SciCodeClaude Opus 5 by 4.352.156.4+4.3
AgentsClaude Opus 5
- τ-Bench V3 · BankingClaude Opus 5 by 3111.142.1+31
- MCP AtlasClaude Opus 5 by 25.260.685.8+25.2
- BrowseCompClaude Opus 5 by 2565.890.8+25
- GDPValClaude Opus 5 by 12.948.361.2+12.9
ReasoningClaude Opus 5
- ARC-AGI-2Claude Opus 5 by 37.552.990.4+37.5
- SimpleBenchClaude Opus 5 by 34.845.880.6+34.8
- CritPtClaude Opus 5 by 17.511.629.1+17.5
- Humanity's Last ExamClaude Opus 5 by 17.237.754.9+17.2
- LiveBench · ReasoningClaude Opus 5 by 883.291.2+8
- GPQA DiamondClaude Opus 5 by 2.990.393.2+2.9
FactsClaude Opus 5
- SimpleQA VerifiedClaude Opus 5 by 22.837.159.9+22.8
- AA-Omniscience · Non-hallucinationClaude Opus 5 by 20.418.839.2+20.4
- AA-Omniscience · AccuracyClaude Opus 5 by 16.644.360.9+16.6
Images and chartsClaude Opus 5
- MMMU-ProClaude Opus 5 by 10.174.684.7+10.1
- LMArena · VisionClaude Opus 5 by 75 rating points12461321+75 rating
- CharXiv (RQ)Claude Opus 5 by 1.682.183.7+1.6
MathClaude Opus 5
- FrontierMath Tier 4Claude Opus 5 by 41.531.773.2+41.5
- FrontierMath Tiers 1-3 (v2)Claude Opus 5 by 18.267.485.6+18.2
- LiveBench · MathematicsClaude Opus 5 by 2.593.295.7+2.5
Following instructionsClaude Opus 5
- LiveBench · Instruction FollowingClaude Opus 5 by 261.863.8+2
Other results11 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.
- AA-OmniscienceClaude Opus 5 by 38-0.937.1+38
- Artificial Analysis Coding IndexClaude Opus 5 by 29.348.778+29.3
- GDPval-AA (Elo)Claude Opus 5 by 246 rating points14621708+246 rating
- AA IntelligenceClaude Opus 5 by 20.430.450.8+20.4
- HLE (with tools)Claude Opus 5 by 18.145.563.6+18.1
- HealthBench Professional length-adjustedClaude Opus 5 by 13.945.959.8+13.9
- ARC-AGI-1Claude Opus 5 by 11.386.297.5+11.3
- livebench_languageClaude Opus 5 by 8.979.888.7+8.9
- AA Agentic IndexGPT-5.2 by 460.256.2+4
- livebench_data_analysisGPT-5.2 by 3.678.274.5+3.6
- HealthBench length-adjustedClaude Opus 5 by 156.857.8+1
Questions people ask
Which is better, GPT-5.2 or Claude Opus 5?
Claude Opus 5 wins seven of the eight areas we test: coding, agents, reasoning, facts, images and charts, math and following instructions. GPT-5.2 wins long documents, and costs 48% less.
Which is better for coding?
Claude Opus 5. It wins 7 of the 7 coding tests both models report; GPT-5.2 wins none.
Which is cheaper?
GPT-5.2 costs $1.75 per million input tokens and $14.00 per million output tokens; Claude Opus 5 costs $5.00 and $25.00. That makes GPT-5.2 about 48% cheaper for the same work.
How do you compare the two?
We use the 39 benchmark tests both models have published scores on. The verdict counts the 28 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 11 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.