GPT-4.1 vs GPT-4o
Wins 6 of 7 areas
Coding · Agents · Reasoning · Facts · Images and charts · Long documents
Wins 0 of 7 areas
—
GPT-4.1 is the stronger all-rounder.
Scores updated · 41 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 software40GPT-4.14 of 4 tests
- FactsGetting facts right instead of making them up31GPT-4.13 of 4 tests
- Images and chartsUnderstanding pictures, charts and video31GPT-4.13 of 4 tests
- ReasoningHard problems that need careful thinking20GPT-4.12 of 3 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own10GPT-4.11 of 1 test
- Long documentsFinding answers in very long texts10GPT-4.11 of 1 test
- Following instructionsDoing exactly what it is asked11Even1 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-4.1 costs 50% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where GPT-4.1 pulls ahead
- Fixes real GitHub issues in Python projectsSWE-bench Verified+21.4points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+14points ahead
- Recent programming contest problemsLiveCodeBench v6+13.8points ahead
Where GPT-4o pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+55.4points ahead
- Reasoning about charts from research papersCharXiv (RQ)+2.1points ahead
- Keeps track of context across a multi-turn chatMulti-Challenge+2points ahead
Every test, side by side
All 41 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGPT-4.1
- SWE-bench VerifiedGPT-4.1 by 21.454.633.2+21.4
- LiveCodeBench v6GPT-4.1 by 13.844.730.9+13.8
- Terminal-Bench HardGPT-4.1 by 5.313.68.3+5.3
- SciCodeGPT-4.1 by 4.838.133.3+4.8
ReasoningGPT-4.1
- GPQA DiamondGPT-4.1 by 1466.652.6+14
- Humanity's Last ExamGPT-4.1 by 2.44.21.8+2.4
- CritPttie00tie
FactsGPT-4.1
- AA-Omniscience · Non-hallucinationGPT-4o by 55.46.762.1+55.4
- SimpleQA VerifiedGPT-4.1 by 5.131.126+5.1
- AA-Omniscience · AccuracyGPT-4.1 by 4.127.823.7+4.1
- Vectara HHEM hallucination ratelower is betterGPT-4.1 by 45.69.6+4
Images and chartsGPT-4.1
- MathVistaGPT-4.1 by 10.872.261.4+10.8
- MMMU-ProGPT-4.1 by 4.961.256.3+4.9
- MMMUGPT-4.1 by 2.674.872.2+2.6
- CharXiv (RQ)GPT-4o by 2.156.758.8+2.1
Following instructionsEven
- IFBenchGPT-4.1 by 74336+7
- Multi-ChallengeGPT-4o by 238.340.3+2
Other results22 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-4.1 by 34.746.411.7+34.7
- AA-OmniscienceGPT-4o by 29.1-39.6-10.5+29.1
- Aider-PolyglotGPT-4.1 by 20.951.630.7+20.9
- AA Agentic IndexGPT-4.1 by 18.927.38.4+18.9
- τ²-Bench Telecom (AA run)GPT-4.1 by 18.247.128.9+18.2
- MMLU-ProXGPT-4.1 by 11.672.761.1+11.6
- HarmBenchGPT-4.1 by 8.891.782.9+8.8
- TAU-bench (retail)GPT-4.1 by 7.76860.3+7.7
- TAU-bench (airline)GPT-4.1 by 6.649.442.8+6.6
- IFEvalGPT-4.1 by 6.487.481+6.4
- vectara_answer_rateGPT-4.1 by 6.199.993.8+6.1
- MMMLUGPT-4.1 by 5.987.381.4+5.9
- AA IntelligenceGPT-4.1 by 5.412.77.3+5.4
- vectara_avg_summary_lengthGPT-4.1 by 5.191.786.6+5.1
- vectara_factual_consistencyGPT-4.1 by 494.490.4+4
- bbqGPT-4o by 2.592.695.1+2.5
- AIR-Bench 2024GPT-4.1 by 2.464.862.4+2.4
- Artificial Analysis Coding IndexGPT-4o by 2.421.824.2+2.4
- simple_safety_testsGPT-4.1 by 1.510098.5+1.5
- ARC-AGI-1GPT-4.1 by 15.54.5+1
- XSTesttie97.997.3tie
- anthropic_red_teamtie99.399.1tie
Questions people ask
Which is better, GPT-4.1 or GPT-4o?
GPT-4.1 wins six of the seven areas where both have results: coding, agents, reasoning, facts, images and charts and long documents. GPT-4o wins none. They are level on following instructions.
Which is better for coding?
GPT-4.1. It wins 4 of the 4 coding tests both models report; GPT-4o wins none.
Which is cheaper?
GPT-4.1 costs $2.00 per million input tokens and $8.00 per million output tokens; GPT-4o costs $5.00 and $15.00. That makes GPT-4.1 about 50% cheaper for the same work.
How do you compare the two?
We use the 41 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 22 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.