Claude 3.5 Sonnet is the stronger all-rounder.GPT-4o mini is cheaper.
Scores updated · 36 tests both models report · How we compare
Where each one wins
Tests won in each of the three 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.
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 mini costs 96% 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 3.5 Sonnet pulls ahead
- Fixes real GitHub issues in Python projectsSWE-bench Verified+40.3points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+13.4points ahead
- Harder college exam questions with imagesMMMU-Pro+13.2points ahead
Where GPT-4o mini pulls ahead
- Very hard expert questions across many subjectsHumanity's Last Exam+1points ahead
Every test, side by side
All 36 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude 3.5 Sonnet
- SWE-bench VerifiedClaude 3.5 Sonnet by 40.3498.7+40.3
- SciCodeClaude 3.5 Sonnet by 8.731.622.9+8.7
ReasoningEven
- GPQA DiamondClaude 3.5 Sonnet by 13.45642.6+13.4
- Humanity's Last ExamGPT-4o mini by 13.24.2+1
Images and chartsClaude 3.5 Sonnet
Other results28 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.
- Aider-PolyglotClaude 3.5 Sonnet by 41.745.33.6+41.7
- AIR-Bench 2024Claude 3.5 Sonnet by 29.685.956.3+29.6
- MMLUClaude 3.5 Sonnet by 21.988.766.8+21.9
- AI2DClaude 3.5 Sonnet by 19.594.775.2+19.5
- SimpleQAClaude 3.5 Sonnet by 18.528.49.9+18.5
- MMMU (val) (Pass@1)Claude 3.5 Sonnet by 16.268.352.1+16.2
- MMLU-ProClaude 3.5 Sonnet by 15.977.661.7+15.9
- Artificial Analysis Coding IndexClaude 3.5 Sonnet by 14.62611.4+14.6
- HarmBenchClaude 3.5 Sonnet by 13.298.184.9+13.2
- GSM8KClaude 3.5 Sonnet by 12.196.484.3+12.1
- naturalquestions_closedbookClaude 3.5 Sonnet by 11.650.238.5+11.6
- ScienceQA (img-test)GPT-4o mini by 10.273.884+10.2
- MATHGPT-4o mini by 9.171.180.2+9.1
- DROPClaude 3.5 Sonnet by 7.487.179.7+7.4
- POPE (test)GPT-4o mini by 776.683.6+7
- HumanEvalClaude 3.5 Sonnet by 6.593.787.2+6.5
- InterGPS (test)Claude 3.5 Sonnet by 5.745.639.9+5.7
- Video-MME OverallGPT-4o mini by 5.355.961.2+5.3
- OpenBookQAClaude 3.5 Sonnet by 5.297.292+5.2
- MGSMClaude 3.5 Sonnet by 4.691.687+4.6
- LiveCodeBenchGPT-4o mini by 2.732.835.5+2.7
- NarrativeQAGPT-4o mini by 2.274.676.8+2.2
- simple_safety_testsClaude 3.5 Sonnet by 2.210097.8+2.2
- anthropic_red_teamClaude 3.5 Sonnet by 1.599.898.3+1.5
- MMBench (dev en)GPT-4o mini by 1.582.383.8+1.5
- AA Intelligencetie7.26.7tie
- TextVQA-valtie70.570.9tie
- XSTesttie95.696tie
Questions people ask
Which is better, Claude 3.5 Sonnet or GPT-4o mini?
Claude 3.5 Sonnet wins two of the three areas where both have results: coding and images and charts. GPT-4o mini wins none, but costs 96% less. They are level on reasoning.
Which is better for coding?
Claude 3.5 Sonnet. It wins 2 of the 2 coding tests both models report; GPT-4o mini wins none.
Which is cheaper?
Claude 3.5 Sonnet costs $3.00 per million input tokens and $15.00 per million output tokens; GPT-4o mini costs $0.15 and $0.60. That makes GPT-4o mini about 96% cheaper for the same work.
How do you compare the two?
We use the 36 benchmark tests both models have published scores on. The verdict counts the 8 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 28 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.