Claude Opus 4 vs GPT-4.5 Preview
Wins 3 of 4 areas
Coding · Reasoning · Following instructions
Wins 1 of 4 areas
Images and charts
Claude Opus 4 is the stronger all-rounder.GPT-4.5 Preview is better at images and charts.
Scores updated · 20 tests both models report · How we compare
Where each one wins
Tests won in each of the four 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.
- ReasoningHard problems that need careful thinking20Claude Opus 42 of 2 tests
- CodingWriting and fixing software10Claude Opus 41 of 1 test
- Following instructionsDoing exactly what it is asked10Claude Opus 41 of 1 test
- Images and chartsUnderstanding pictures, charts and video01GPT-4.5 Preview1 of 2 tests · 1 tie
Coding 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.
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+34.5points ahead
- Common-sense trick questionsSimpleBench+24.3points ahead
- Keeps track of context across a multi-turn chatMulti-Challenge+14.8points ahead
Where GPT-4.5 Preview pulls ahead
- College exam questions with charts, maps and diagramsMMMU+1.5points ahead
Every test, side by side
All 20 tests both models report. The winning score is in its model's colour; marks a score checked independently.
ReasoningClaude Opus 4
- SimpleBenchClaude Opus 4 by 24.358.834.5+24.3
- GPQA DiamondClaude Opus 4 by 10.179.669.5+10.1
Images and chartsGPT-4.5 Preview
- MMMUGPT-4.5 Preview by 1.573.775.2+1.5
- LMArena · Visiontie11921195tie
Following instructionsClaude Opus 4
- Multi-ChallengeClaude Opus 4 by 14.858.643.8+14.8
Other results14 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.
- SimpleQAGPT-4.5 Preview by 39.722.862.5+39.7
- Aider-PolyglotClaude Opus 4 by 25.870.744.9+25.8
- ARC-AGI-1Claude Opus 4 by 25.435.710.3+25.4
- TAU-bench (retail)Claude Opus 4 by 1381.468.4+13
- AIR-Bench 2024Claude Opus 4 by 11.585.774.2+11.5
- AA IntelligenceClaude Opus 4 by 1120.69.6+11
- TAU-bench (airline)Claude Opus 4 by 9.659.650+9.6
- bbqClaude Opus 4 by 5.197.192+5.1
- HarmBenchGPT-4.5 Preview by 4.191.795.8+4.1
- MMMLUClaude Opus 4 by 3.788.885.1+3.7
- XSTestClaude Opus 4 by 1.99795.1+1.9
- IFEvaltie87.488.2tie
- anthropic_red_teamtie98.999.4tie
- simple_safety_teststie100100tie
Questions people ask
Which is better, Claude Opus 4 or GPT-4.5 Preview?
Claude Opus 4 wins three of the four areas where both have results: coding, reasoning and following instructions. GPT-4.5 Preview wins images and charts.
Which is better for coding?
Claude Opus 4. It wins the one coding test both models report.
How do you compare the two?
We use the 20 benchmark tests both models have published scores on. The verdict counts the 6 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 14 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.