Claude Opus 4 vs GPT-4o
Wins 5 of 6 areas
Coding · Reasoning · Images and charts · Long documents · Following instructions
Wins 1 of 6 areas
Facts
Claude Opus 4 is the stronger all-rounder.GPT-4o is cheaper and better at facts.
Scores updated · 48 tests both models report · How we compare
Where each one wins
Tests won in each of the six areas where both have results. 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 software40Claude Opus 44 of 4 tests
- ReasoningHard problems that need careful thinking30Claude Opus 43 of 3 tests
- Long documentsFinding answers in very long texts20Claude Opus 42 of 2 tests
- Following instructionsDoing exactly what it is asked20Claude Opus 42 of 2 tests
- Images and chartsUnderstanding pictures, charts and video10Claude Opus 41 of 1 test
- FactsGetting facts right instead of making them up01GPT-4o1 of 1 test
Facts and images and charts 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-4o costs 78% 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 4 pulls ahead
- Fixes real GitHub issues in Python projectsSWE-bench Verified+39.3points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+27points ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+22.8points ahead
Where GPT-4o pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 48 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Opus 4
- SWE-bench VerifiedClaude Opus 4 by 39.372.533.2+39.3
- Terminal-Bench HardClaude Opus 4 by 22.831.18.3+22.8
- LiveCodeBench v6Claude Opus 4 by 16.547.430.9+16.5
- SciCodeClaude Opus 4 by 6.539.833.3+6.5
ReasoningClaude Opus 4
- GPQA DiamondClaude Opus 4 by 2779.652.6+27
- Humanity's Last ExamClaude Opus 4 by 10.512.31.8+10.5
- ARC-AGI-2Claude Opus 4 by 8.68.60+8.6
Long documentsClaude Opus 4
- AA-LCRClaude Opus 4 by 13.369.356+13.3
- LongBench v2Claude Opus 4 by 7.555.648.1+7.5
Following instructionsClaude Opus 4
- Multi-ChallengeClaude Opus 4 by 18.358.640.3+18.3
- IFBenchClaude Opus 4 by 17.753.736+17.7
Other results35 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.
- AIME 2025Claude Opus 4 by 63.875.511.7+63.8
- livecodebench_mediumClaude Opus 4 by 48.380.432.1+48.3
- CNMO 2024Claude Opus 4 by 46.857.610.8+46.8
- τ²-Bench Telecom (AA run)Claude Opus 4 by 44.573.428.9+44.5
- Aider-PolyglotClaude Opus 4 by 4070.730.7+40
- AIME 2024Claude Opus 4 by 38.948.29.3+38.9
- LiveCodeBenchClaude Opus 4 by 32.170.438.3+32.1
- ARC-AGI-1Claude Opus 4 by 31.235.74.5+31.2
- livecodebench_hardClaude Opus 4 by 28.633.14.5+28.6
- AIR-Bench 2024Claude Opus 4 by 23.385.762.4+23.3
- TAU-bench (retail)Claude Opus 4 by 21.181.460.3+21.1
- metr_re_benchClaude Opus 4 by 20200+20
- MATH-500 (EM)Claude Opus 4 by 19.894.474.6+19.8
- MMLUClaude Opus 4 by 19.192.973.8+19.1
- τ²-Bench (Retail)Claude Opus 4 by 18.481.863.4+18.4
- TAU-bench (airline)Claude Opus 4 by 16.859.642.8+16.8
- livecodebench_easyClaude Opus 4 by 16.699.182.5+16.6
- SimpleQAGPT-4o by 15.422.838.2+15.4
- Tau2 airlineClaude Opus 4 by 14.56045.5+14.5
- SuperGPQAClaude Opus 4 by 14.156.542.4+14.1
- AA IntelligenceClaude Opus 4 by 13.320.67.3+13.3
- MMLU-ProClaude Opus 4 by 11.986.674.7+11.9
- Artificial Analysis Coding IndexClaude Opus 4 by 9.83424.2+9.8
- HarmBenchClaude Opus 4 by 8.891.782.9+8.8
- MMMLUClaude Opus 4 by 7.488.881.4+7.4
- IFEvalClaude Opus 4 by 6.487.481+6.4
- MMLU-ReduxClaude Opus 4 by 6.294.288+6.2
- vectara_avg_summary_lengthClaude Opus 4 by 36.6 rating points123.286.6+36.6 rating
- vectara_answer_rateGPT-4o by 2.89193.8+2.8
- vectara_factual_consistencyGPT-4o by 2.48890.4+2.4
- bbqClaude Opus 4 by 297.195.1+2
- simple_safety_testsClaude Opus 4 by 1.510098.5+1.5
- metr_swaatie99.899.2tie
- XSTesttie9797.3tie
- anthropic_red_teamtie98.999.1tie
Questions people ask
Which is better, Claude Opus 4 or GPT-4o?
Claude Opus 4 wins five of the six areas where both have results: coding, reasoning, images and charts, long documents and following instructions. GPT-4o wins facts, and costs 78% less.
Which is better for coding?
Claude Opus 4. It wins 4 of the 4 coding tests both models report; GPT-4o wins none.
Which is cheaper?
Claude Opus 4 costs $15.00 per million input tokens and $75.00 per million output tokens; GPT-4o costs $5.00 and $15.00. That makes GPT-4o about 78% cheaper for the same work.
How do you compare the two?
We use the 48 benchmark tests both models have published scores on. The verdict counts the 13 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 35 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.