Claude Sonnet 4.5 vs GPT-5 nano
Wins 7 of 8 areas
Coding · Agents · Reasoning · Facts · Images and charts · Math · Long documents
Wins 1 of 8 areas
Following instructions
Claude Sonnet 4.5 is the stronger all-rounder.GPT-5 nano is cheaper and better at following instructions.
Scores updated · 37 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 thinking40Claude Sonnet 4.54 of 4 tests
- CodingWriting and fixing software30Claude Sonnet 4.53 of 3 tests
- FactsGetting facts right instead of making them up31Claude Sonnet 4.53 of 4 tests
- MathCompetition and research-level math10Claude Sonnet 4.51 of 2 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own10Claude Sonnet 4.51 of 1 test
- Images and chartsUnderstanding pictures, charts and video10Claude Sonnet 4.51 of 1 test
- Long documentsFinding answers in very long texts10Claude Sonnet 4.51 of 1 test
- Following instructionsDoing exactly what it is asked01GPT-5 nano1 of 1 test
Agents, images and charts, 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 nano costs 98% 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
- Fixes real GitHub issues in Python projectsSWE-bench Verified+42.4points ahead
- Reasons across sets of long documentsAA-LCR+27.3points ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+23.5points ahead
Where GPT-5 nano pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+10.3points ahead
Every test, side by side
All 37 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 42.477.234.8+42.4
- Terminal-Bench HardClaude Sonnet 4.5 by 23.535.612.1+23.5
- SciCodeClaude Sonnet 4.5 by 9.145.736.6+9.1
ReasoningClaude Sonnet 4.5
- GPQA DiamondClaude Sonnet 4.5 by 15.883.467.6+15.8
- ARC-AGI-2Claude Sonnet 4.5 by 1113.62.6+11
- Humanity's Last ExamClaude Sonnet 4.5 by 8.317.89.5+8.3
- CritPtClaude Sonnet 4.5 by 1.11.10+1.1
FactsClaude Sonnet 4.5
- SimpleQA VerifiedClaude Sonnet 4.5 by 1930.711.7+19
- AA-Omniscience · AccuracyClaude Sonnet 4.5 by 13.832.919.1+13.8
- AA-Omniscience · Non-hallucinationClaude Sonnet 4.5 by 9.950.941+9.9
- Vectara HHEM hallucination ratelower is betterGPT-5 nano by 1.51210.5+1.5
Images and chartsClaude Sonnet 4.5
- MMMU-ProClaude Sonnet 4.5 by 7.768.761+7.7
MathClaude Sonnet 4.5
- FrontierMath Tiers 1-3 (v2)Claude Sonnet 4.5 by 3.923.920+3.9
- FrontierMath Tier 4tie2.42.4tie
Long documentsClaude Sonnet 4.5
- AA-LCRClaude Sonnet 4.5 by 27.372.345+27.3
Following instructionsGPT-5 nano
- IFBenchGPT-5 nano by 10.357.367.6+10.3
Other results20 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.
- ARC-AGI-1Claude Sonnet 4.5 by 4763.716.7+47
- τ²-Bench Telecom (AA run)Claude Sonnet 4.5 by 41.678.136.5+41.6
- swe_bench_bashClaude Sonnet 4.5 by 36.671.434.8+36.6
- Artificial Analysis Coding IndexClaude Sonnet 4.5 by 31.852.120.3+31.8
- AA-OmniscienceClaude Sonnet 4.5 by 28.6-0.1-28.7+28.6
- AA IntelligenceClaude Sonnet 4.5 by 7.720.713+7.7
- HarmBenchGPT-5 nano by 6.49298.4+6.4
- AA Agentic IndexGPT-5 nano by 5.717.523.2+5.7
- vectara_answer_rateGPT-5 nano by 4.495.6100+4.4
- vectara_avg_summary_lengthClaude Sonnet 4.5 by 22.1 rating points127.8105.7+22.1 rating
- frontiermath_tier_4_v1Claude Sonnet 4.5 by 2.14.22.1+2.1
- AIR-Bench 2024Claude Sonnet 4.5 by 289.887.8+2
- AIME 2025Claude Sonnet 4.5 by 1.88785.2+1.8
- vectara_factual_consistencyGPT-5 nano by 1.58889.5+1.5
- anthropic_red_teamGPT-5 nano by 1.498.299.5+1.4
- bbqClaude Sonnet 4.5 by 1.398.997.6+1.3
- XSTestGPT-5 nano by 1.296.497.6+1.2
- HMMT 2025GPT-5 nano by 174.675.6+1
- HMMT Nov. 2025tie81.781.7tie
- simple_safety_teststie100100tie
Questions people ask
Which is better, Claude Sonnet 4.5 or GPT-5 nano?
Claude Sonnet 4.5 wins seven of the eight areas we test: coding, agents, reasoning, facts, images and charts, math and long documents. GPT-5 nano wins following instructions, and costs 98% less.
Which is better for coding?
Claude Sonnet 4.5. It wins 3 of the 3 coding tests both models report; GPT-5 nano wins none.
Which is cheaper?
Claude Sonnet 4.5 costs $3.00 per million input tokens and $15.00 per million output tokens; GPT-5 nano costs $0.05 and $0.40. That makes GPT-5 nano about 98% cheaper for the same work.
How do you compare the two?
We use the 37 benchmark tests both models have published scores on. The verdict counts the 17 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 20 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.