GPT-5 vs GPT-5 mini
Wins 6 of 8 areas
Coding · Agents · Reasoning · Images and charts · Math · Long documents
Wins 0 of 8 areas
—
GPT-5 is the stronger all-rounder.GPT-5 mini is cheaper.
Scores updated · 44 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 thinking40GPT-54 of 4 tests
- CodingWriting and fixing software30GPT-53 of 4 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own20GPT-52 of 2 tests
- MathCompetition and research-level math20GPT-52 of 2 tests
- Images and chartsUnderstanding pictures, charts and video10GPT-51 of 2 tests · 1 tie
- Long documentsFinding answers in very long texts10GPT-51 of 1 test
- FactsGetting facts right instead of making them up22Even2 each
- Following instructionsDoing exactly what it is asked11Even1 each
Long documents rests on a single test.
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 mini costs 80% 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 GPT-5 pulls ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+31.5points ahead
- Short factual questions, answered correctlySimpleQA Verified+28.5points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+15.3points ahead
Where GPT-5 mini pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+25.8points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+2.3points ahead
Every test, side by side
All 44 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGPT-5
- Terminal-Bench 2.1GPT-5 by 31.535.23.7+31.5
- SWE-bench VerifiedGPT-5 by 15.174.959.8+15.1
- SciCodeGPT-5 by 3.942.939+3.9
- Terminal-Bench Hardtie32.633.3tie
AgentsGPT-5
- GDPValGPT-5 by 7.52113.5+7.5
- τ-Bench V3 · BankingGPT-5 by 6.622.115.5+6.6
ReasoningGPT-5
- Humanity's Last ExamGPT-5 by 728.521.5+7
- CritPtGPT-5 by 5.75.70+5.7
- ARC-AGI-2GPT-5 by 5.59.94.4+5.5
- GPQA DiamondGPT-5 by 2.685.482.8+2.6
FactsEven
- SimpleQA VerifiedGPT-5 by 28.550.121.6+28.5
- AA-Omniscience · Non-hallucinationGPT-5 mini by 25.817.843.6+25.8
- AA-Omniscience · AccuracyGPT-5 by 15.340.325+15.3
- Vectara HHEM hallucination ratelower is betterGPT-5 mini by 2.215.112.9+2.2
Images and chartsGPT-5
- MMMU-ProGPT-5 by 4.174.270.1+4.1
- LMArena · Visiontie12081202tie
MathGPT-5
- FrontierMath Tier 4GPT-5 by 9.821.912.2+9.8
- FrontierMath Tiers 1-3 (v2)GPT-5 by 8.755.446.7+8.7
Following instructionsEven
- Multi-ChallengeGPT-5 by 10.669.659+10.6
- IFBenchGPT-5 mini by 2.373.175.4+2.3
Other results23 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.
- Artificial Analysis Coding IndexGPT-5 by 22.237.815.6+22.2
- AA Agentic IndexGPT-5 by 17.626.58.9+17.6
- τ²-Bench Telecom (AA run)GPT-5 by 16.484.868.4+16.4
- ARC-AGI-1GPT-5 by 11.465.754.3+11.4
- LiveCodeBench (Jan-May)GPT-5 by 9.486.877.4+9.4
- AA-OmniscienceGPT-5 by 8.6-8.7-17.3+8.6
- AA IntelligenceGPT-5 by 6.22316.8+6.2
- frontiermath_tier_4_v1GPT-5 by 6.212.56.3+6.2
- HMMT 2025GPT-5 by 5.593.387.8+5.5
- swe_bench_bashGPT-5 by 5.26559.8+5.2
- HMMT Nov. 2025GPT-5 by 589.284.2+5
- FrontierMath (overall)GPT-5 by 4.226.322.1+4.2
- AIME 2025GPT-5 by 3.594.691.1+3.5
- vectara_factual_consistencyGPT-5 mini by 2.284.987.1+2.2
- AIR-Bench 2024GPT-5 by 287.785.7+2
- vectara_avg_summary_lengthtie162.7169.7tie
- matharena_visual_math_overalltie78.878.2tie
- XSTesttie97.197.7tie
- bbqtie96.896.3tie
- HarmBenchtie97.697.1tie
- simple_safety_teststie99.8100tie
- anthropic_red_teamtie99.199.1tie
- vectara_answer_ratetie99.999.9tie
Questions people ask
Which is better, GPT-5 or GPT-5 mini?
GPT-5 wins six of the eight areas we test: coding, agents, reasoning, images and charts, math and long documents. GPT-5 mini wins none, but costs 80% less. They are level on facts and following instructions.
Which is better for coding?
GPT-5. It wins 3 of the 4 coding tests both models report; GPT-5 mini wins none, and 1 is a tie.
Which is cheaper?
GPT-5 costs $1.25 per million input tokens and $10.00 per million output tokens; GPT-5 mini costs $0.25 and $2.00. That makes GPT-5 mini about 80% cheaper for the same work.
How do you compare the two?
We use the 44 benchmark tests both models have published scores on. The verdict counts the 21 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 23 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.