GPT-4o vs GPT-5
Wins 0 of 7 areas
—
Wins 6 of 7 areas
Coding · Agents · Reasoning · Images and charts · Long documents · Following instructions
GPT-5 is the stronger all-rounder.
Scores updated · 46 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 software04GPT-54 of 4 tests
- ReasoningHard problems that need careful thinking04GPT-54 of 4 tests
- Images and chartsUnderstanding pictures, charts and video03GPT-53 of 3 tests
- Following instructionsDoing exactly what it is asked02GPT-52 of 2 tests
- AgentsCarrying out multi-step tasks on its own01GPT-51 of 1 test
- Long documentsFinding answers in very long texts01GPT-51 of 1 test
- FactsGetting facts right instead of making them up22Even2 each
Agents and long documents 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 costs 44% 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
- Recent programming contest problemsLiveCodeBench v6+56.1points ahead
- Fixes real GitHub issues in Python projectsSWE-bench Verified+41.7points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+37.1points ahead
Where GPT-4o pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+44.3points ahead
Every test, side by side
All 46 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGPT-5
- LiveCodeBench v6GPT-5 by 56.130.987+56.1
- SWE-bench VerifiedGPT-5 by 41.733.274.9+41.7
- Terminal-Bench HardGPT-5 by 24.38.332.6+24.3
- SciCodeGPT-5 by 9.633.342.9+9.6
ReasoningGPT-5
- GPQA DiamondGPT-5 by 32.852.685.4+32.8
- Humanity's Last ExamGPT-5 by 26.71.828.5+26.7
- ARC-AGI-2GPT-5 by 9.909.9+9.9
- CritPtGPT-5 by 5.705.7+5.7
FactsEven
- AA-Omniscience · Non-hallucinationGPT-4o by 44.362.117.8+44.3
- SimpleQA VerifiedGPT-5 by 24.12650.1+24.1
- AA-Omniscience · AccuracyGPT-5 by 16.623.740.3+16.6
- Vectara HHEM hallucination ratelower is betterGPT-4o by 5.59.615.1+5.5
Images and chartsGPT-5
- CharXiv (RQ)GPT-5 by 22.358.881.1+22.3
- MMMU-ProGPT-5 by 17.956.374.2+17.9
- MMMUGPT-5 by 1272.284.2+12
Following instructionsGPT-5
- IFBenchGPT-5 by 37.13673.1+37.1
- Multi-ChallengeGPT-5 by 29.340.369.6+29.3
Other results27 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 2025GPT-5 by 82.911.794.6+82.9
- ARC-AGI-1GPT-5 by 61.24.565.7+61.2
- Aider-PolyglotGPT-5 by 57.330.788+57.3
- τ²-Bench Telecom (AA run)GPT-5 by 55.928.984.8+55.9
- LiveCodeBenchGPT-5 by 45.238.383.5+45.2
- metr_hcastGPT-5 by 43.42568.4+43.4
- metr_re_benchGPT-5 by 40040+40
- ERQAGPT-5 by 30.535.265.7+30.5
- AIR-Bench 2024GPT-5 by 25.362.487.7+25.3
- VideoMMMUGPT-5 by 23.461.284.6+23.4
- AA Agentic IndexGPT-5 by 18.18.426.5+18.1
- τ²-Bench (Retail)GPT-5 by 17.763.481.1+17.7
- Tau2 airlineGPT-5 by 17.145.562.6+17.1
- AA IntelligenceGPT-5 by 15.77.323+15.7
- HarmBenchGPT-5 by 14.782.997.6+14.7
- Artificial Analysis Coding IndexGPT-5 by 13.624.237.8+13.6
- MMLU-ProGPT-5 by 12.474.787.1+12.4
- vectara_avg_summary_lengthGPT-5 by 76.1 rating points86.6162.7+76.1 rating
- MMLU-ReduxGPT-5 by 7.38895.3+7.3
- vectara_answer_rateGPT-5 by 6.193.899.9+6.1
- vectara_factual_consistencyGPT-4o by 5.590.484.9+5.5
- HumanEvalGPT-5 by 3.290.293.4+3.2
- AA-OmniscienceGPT-5 by 1.8-10.5-8.7+1.8
- bbqGPT-5 by 1.795.196.8+1.7
- simple_safety_testsGPT-5 by 1.398.599.8+1.3
- XSTesttie97.397.1tie
- anthropic_red_teamtie99.199.1tie
Questions people ask
Which is better, GPT-4o or GPT-5?
GPT-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?
GPT-5. It wins 4 of the 4 coding tests both models report; GPT-4o wins none.
Which is cheaper?
GPT-4o costs $5.00 per million input tokens and $15.00 per million output tokens; GPT-5 costs $1.25 and $10.00. That makes GPT-5 about 44% cheaper for the same work.
How do you compare the two?
We use the 46 benchmark tests both models have published scores on. The verdict counts the 19 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 27 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.