GPT-4.1 vs GPT-4.1 nano
Wins 7 of 7 areas
Coding · Agents · Reasoning · Facts · Images and charts · Long documents · Following instructions
Wins 0 of 7 areas
—
GPT-4.1 is the stronger all-rounder.GPT-4.1 nano is cheaper.
Scores updated · 32 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.
- Images and chartsUnderstanding pictures, charts and video40GPT-4.14 of 4 tests
- CodingWriting and fixing software20GPT-4.12 of 2 tests
- Following instructionsDoing exactly what it is asked20GPT-4.12 of 2 tests
- ReasoningHard problems that need careful thinking10GPT-4.11 of 3 tests · 2 ties
- FactsGetting facts right instead of making them up21GPT-4.12 of 3 tests
- 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
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 nano costs 95% 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-4.1 pulls ahead
- Reasons across sets of long documentsAA-LCR+48points ahead
- Short factual questions, answered correctlySimpleQA Verified+25.1points ahead
- Keeps track of context across a multi-turn chatMulti-Challenge+23.3points ahead
Where GPT-4.1 nano pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+10.7points ahead
Every test, side by side
All 32 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGPT-4.1
- SciCodeGPT-4.1 by 12.238.125.9+12.2
- Terminal-Bench HardGPT-4.1 by 9.813.63.8+9.8
ReasoningGPT-4.1
- GPQA DiamondGPT-4.1 by 15.466.651.2+15.4
- Humanity's Last Examtie4.23.8tie
- CritPttie00tie
FactsGPT-4.1
- SimpleQA VerifiedGPT-4.1 by 25.131.16+25.1
- AA-Omniscience · AccuracyGPT-4.1 by 14.127.813.7+14.1
- AA-Omniscience · Non-hallucinationGPT-4.1 nano by 10.76.717.4+10.7
Images and chartsGPT-4.1
- MMMU-ProGPT-4.1 by 21.161.240.1+21.1
- MMMUGPT-4.1 by 19.474.855.4+19.4
- CharXiv (RQ)GPT-4.1 by 16.256.740.5+16.2
- MathVistaGPT-4.1 by 1672.256.2+16
Following instructionsGPT-4.1
- Multi-ChallengeGPT-4.1 by 23.338.315+23.3
- IFBenchGPT-4.1 by 114332+11
Other results16 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.
- TAU-bench (retail)GPT-4.1 by 45.46822.6+45.4
- Aider-PolyglotGPT-4.1 by 41.851.69.8+41.8
- TAU-bench (airline)GPT-4.1 by 35.449.414+35.4
- τ²-Bench Telecom (AA run)GPT-4.1 by 29.847.117.3+29.8
- MMMLUGPT-4.1 by 20.487.366.9+20.4
- AA-OmniscienceGPT-4.1 by 18-39.6-57.6+18
- IFEvalGPT-4.1 by 12.987.474.5+12.9
- Artificial Analysis Coding IndexGPT-4.1 by 10.721.811.1+10.7
- ARC-AGI-1GPT-4.1 by 5.55.50+5.5
- bbqGPT-4.1 by 5.192.687.5+5.1
- AA IntelligenceGPT-4.1 by 4.912.77.8+4.9
- HarmBenchGPT-4.1 by 4.991.786.8+4.9
- AIR-Bench 2024GPT-4.1 by 3.364.861.5+3.3
- XSTestGPT-4.1 by 1.997.996+1.9
- simple_safety_testsGPT-4.1 by 110099+1
- anthropic_red_teamtie99.399.6tie
Questions people ask
Which is better, GPT-4.1 or GPT-4.1 nano?
GPT-4.1 wins all seven areas where both have results: coding, agents, reasoning, facts, images and charts, long documents and following instructions. GPT-4.1 nano wins none, but costs 95% less.
Which is better for coding?
GPT-4.1. It wins 2 of the 2 coding tests both models report; GPT-4.1 nano wins none.
Which is cheaper?
GPT-4.1 costs $2.00 per million input tokens and $8.00 per million output tokens; GPT-4.1 nano costs $0.10 and $0.40. That makes GPT-4.1 nano about 95% cheaper for the same work.
How do you compare the two?
We use the 32 benchmark tests both models have published scores on. The verdict counts the 16 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 16 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.