Claude Sonnet 4.5 vs GPT-4o
Wins 6 of 7 areas
Coding · Agents · Reasoning · Images and charts · Long documents · Following instructions
Wins 0 of 7 areas
—
Claude Sonnet 4.5 is the stronger all-rounder.
Scores updated · 48 tests both models report · How we compare
Where each one wins
Tests won in each of the seven 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 Sonnet 4.54 of 4 tests
- ReasoningHard problems that need careful thinking40Claude Sonnet 4.54 of 4 tests
- Images and chartsUnderstanding pictures, charts and video40Claude Sonnet 4.54 of 4 tests
- AgentsCarrying out multi-step tasks on its own20Claude Sonnet 4.52 of 2 tests
- Long documentsFinding answers in very long texts20Claude Sonnet 4.52 of 2 tests
- Following instructionsDoing exactly what it is asked20Claude Sonnet 4.52 of 2 tests
- FactsGetting facts right instead of making them up22Even2 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.
Claude Sonnet 4.5 costs 10% 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 4.5 pulls ahead
- Completes tasks by operating a computer desktopOSWorld-Verified+56.4points ahead
- Fixes real GitHub issues in Python projectsSWE-bench Verified+44points ahead
- Recent programming contest problemsLiveCodeBench v6+33.1points ahead
Where GPT-4o pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+11.2points ahead
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 Sonnet 4.5
- SWE-bench VerifiedClaude Sonnet 4.5 by 4477.233.2+44
- LiveCodeBench v6Claude Sonnet 4.5 by 33.16430.9+33.1
- Terminal-Bench HardClaude Sonnet 4.5 by 27.335.68.3+27.3
- SciCodeClaude Sonnet 4.5 by 12.445.733.3+12.4
AgentsClaude Sonnet 4.5
- OSWorld-VerifiedClaude Sonnet 4.5 by 56.461.45+56.4
- GDPValClaude Sonnet 4.5 by 20.520.50+20.5
ReasoningClaude Sonnet 4.5
- GPQA DiamondClaude Sonnet 4.5 by 30.883.452.6+30.8
- Humanity's Last ExamClaude Sonnet 4.5 by 1617.81.8+16
- ARC-AGI-2Claude Sonnet 4.5 by 13.613.60+13.6
- CritPtClaude Sonnet 4.5 by 1.11.10+1.1
FactsEven
- AA-Omniscience · Non-hallucinationGPT-4o by 11.250.962.1+11.2
- AA-Omniscience · AccuracyClaude Sonnet 4.5 by 9.232.923.7+9.2
- SimpleQA VerifiedClaude Sonnet 4.5 by 4.730.726+4.7
- Vectara HHEM hallucination ratelower is betterGPT-4o by 2.4129.6+2.4
Images and chartsClaude Sonnet 4.5
- MathVistaClaude Sonnet 4.5 by 18.479.861.4+18.4
- MMMU-ProClaude Sonnet 4.5 by 12.468.756.3+12.4
- CharXiv (RQ)Claude Sonnet 4.5 by 8.467.258.8+8.4
- MMMUClaude Sonnet 4.5 by 7.479.672.2+7.4
Long documentsClaude Sonnet 4.5
- AA-LCRClaude Sonnet 4.5 by 16.372.356+16.3
- LongBench v2Claude Sonnet 4.5 by 13.761.848.1+13.7
Following instructionsClaude Sonnet 4.5
- IFBenchClaude Sonnet 4.5 by 21.357.336+21.3
- Multi-ChallengeClaude Sonnet 4.5 by 1555.340.3+15
Other results26 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 Sonnet 4.5 by 75.38711.7+75.3
- ARC-AGI-1Claude Sonnet 4.5 by 59.263.74.5+59.2
- τ²-Bench Telecom (AA run)Claude Sonnet 4.5 by 49.278.128.9+49.2
- LiveCodeBenchClaude Sonnet 4.5 by 32.77138.3+32.7
- Artificial Analysis Coding IndexClaude Sonnet 4.5 by 27.952.124.2+27.9
- AIR-Bench 2024Claude Sonnet 4.5 by 27.489.862.4+27.4
- TAU-bench (airline)Claude Sonnet 4.5 by 27.27042.8+27.2
- TAU-bench (retail)Claude Sonnet 4.5 by 25.986.260.3+25.9
- τ²-Bench (Retail)Claude Sonnet 4.5 by 22.886.263.4+22.8
- AA IntelligenceClaude Sonnet 4.5 by 13.420.77.3+13.4
- MMLU-ProClaude Sonnet 4.5 by 12.887.574.7+12.8
- AA-OmniscienceClaude Sonnet 4.5 by 10.4-0.1-10.5+10.4
- AA Agentic IndexClaude Sonnet 4.5 by 9.117.58.4+9.1
- HarmBenchClaude Sonnet 4.5 by 9.19282.9+9.1
- MMMU (val) (Pass@1)Claude Sonnet 4.5 by 8.777.869.1+8.7
- MMMLUClaude Sonnet 4.5 by 7.789.181.4+7.7
- MMLU-ReduxClaude Sonnet 4.5 by 7.695.688+7.6
- AI2DGPT-4o by 7.28794.2+7.2
- RealWorldQAGPT-4o by 5.170.375.4+5.1
- vectara_avg_summary_lengthClaude Sonnet 4.5 by 41.2 rating points127.886.6+41.2 rating
- bbqClaude Sonnet 4.5 by 3.898.995.1+3.8
- vectara_factual_consistencyGPT-4o by 2.48890.4+2.4
- vectara_answer_rateClaude Sonnet 4.5 by 1.895.693.8+1.8
- simple_safety_testsClaude Sonnet 4.5 by 1.510098.5+1.5
- anthropic_red_teamtie98.299.1tie
- XSTesttie96.497.3tie
Questions people ask
Which is better, Claude Sonnet 4.5 or GPT-4o?
Claude Sonnet 4.5 wins six of the seven areas where both have results: coding, agents, reasoning, images and charts, long documents and following instructions. GPT-4o wins none. They are level on facts.
Which is better for coding?
Claude Sonnet 4.5. It wins 4 of the 4 coding tests both models report; GPT-4o wins none.
Which is cheaper?
Claude Sonnet 4.5 costs $3.00 per million input tokens and $15.00 per million output tokens; GPT-4o costs $5.00 and $15.00. That makes Claude Sonnet 4.5 about 10% 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 22 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 26 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.