GPT-5.4 vs GPT-5.4 nano
Wins 6 of 8 areas
Coding · Agents · Reasoning · Images and charts · Math · Long documents
Wins 0 of 8 areas
—
GPT-5.4 is the stronger all-rounder.GPT-5.4 nano is cheaper.
Scores updated · 45 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.
- CodingWriting and fixing software60GPT-5.46 of 6 tests
- AgentsCarrying out multi-step tasks on its own60GPT-5.46 of 6 tests
- ReasoningHard problems that need careful thinking50GPT-5.45 of 5 tests
- MathCompetition and research-level math30GPT-5.43 of 3 tests
- Images and chartsUnderstanding pictures, charts and video20GPT-5.42 of 2 tests
- Long documentsFinding answers in very long texts10GPT-5.41 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-5.4 nano costs 92% 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-5.4 pulls ahead
- Abstract visual puzzles that people can solveARC-AGI-2+67.6points ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+36.8points ahead
- Completes tasks by operating a computer desktopOSWorld-Verified+36points ahead
Where GPT-5.4 nano pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+17.5points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+2points ahead
Every test, side by side
All 45 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGPT-5.4
- Terminal-Bench 2.1GPT-5.4 by 17.678.360.7+17.6
- Terminal-Bench HardGPT-5.4 by 15.257.642.4+15.2
- SciCodeGPT-5.4 by 9.456.647.2+9.4
- LiveBench · Agentic CodingGPT-5.4 by 753.846.8+7
- LiveBench · CodingGPT-5.4 by 6.777.570.8+6.7
- SWE-bench ProGPT-5.4 by 5.357.752.4+5.3
AgentsGPT-5.4
- OSWorld-VerifiedGPT-5.4 by 367539+36
- GDPValGPT-5.4 by 14.937.422.5+14.9
- τ-Bench V3 · BankingGPT-5.4 by 12.239.627.4+12.2
- MCP AtlasGPT-5.4 by 11.167.256.1+11.1
- AA IT-Bench SREGPT-5.4 by 10.134.524.4+10.1
- AA ApexAgentsGPT-5.4 by 8.433.324.9+8.4
ReasoningGPT-5.4
- ARC-AGI-2GPT-5.4 by 67.673.35.7+67.6
- Humanity's Last ExamGPT-5.4 by 15.443.728.3+15.4
- CritPtGPT-5.4 by 14.123.49.3+14.1
- GPQA DiamondGPT-5.4 by 10.39281.7+10.3
- LiveBench · ReasoningGPT-5.4 by 788.181.1+7
FactsEven
- SimpleQA VerifiedGPT-5.4 by 33.445.111.7+33.4
- AA-Omniscience · AccuracyGPT-5.4 by 25.250.925.7+25.2
- AA-Omniscience · Non-hallucinationGPT-5.4 nano by 17.58.325.8+17.5
- Vectara HHEM hallucination ratelower is betterGPT-5.4 nano by 3.973.1+3.9
Images and chartsGPT-5.4
- MMMU-ProGPT-5.4 by 1378.465.4+13
- LMArena · VisionGPT-5.4 by 106 rating points13021196+106 rating
MathGPT-5.4
- FrontierMath Tier 4GPT-5.4 by 36.84912.2+36.8
- FrontierMath Tiers 1-3 (v2)GPT-5.4 by 33.778.644.9+33.7
- LiveBench · MathematicsGPT-5.4 by 3.294.291+3.2
Following instructionsEven
- LiveBench · Instruction FollowingGPT-5.4 by 370.267.2+3
- IFBenchGPT-5.4 nano by 273.975.9+2
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.
- ARC-AGI-1GPT-5.4 by 42.293.751.5+42.2
- AA-OmniscienceGPT-5.4 by 35.35.8-29.5+35.3
- Terminal-Bench 2.0GPT-5.4 by 28.875.146.3+28.8
- AA Agentic IndexGPT-5.4 by 26.544.217.7+26.5
- frontiermath_tier_4_v1GPT-5.4 by 20.827.16.3+20.8
- livebench_languageGPT-5.4 by 20.182.662.5+20.1
- ToolathlonGPT-5.4 by 19.154.635.5+19.1
- AA IntelligenceGPT-5.4 by 18.33920.7+18.3
- Artificial Analysis Coding IndexGPT-5.4 by 1571.156.1+15
- OmniDocBench 1.5GPT-5.4 by 13.389.175.8+13.3
- livebench_data_analysisGPT-5.4 by 11.779.367.6+11.7
- τ²-Bench Telecom (AA run)GPT-5.4 by 11.187.176+11.1
- LiveBenchGPT-5.4 by 10.280.370.1+10.2
- vectara_avg_summary_lengthGPT-5.4 nano by 62.7 rating points81.7144.4+62.7 rating
- vectara_factual_consistencyGPT-5.4 nano by 3.99396.9+3.9
- vectara_answer_ratetie99.9100tie
Questions people ask
Which is better, GPT-5.4 or GPT-5.4 nano?
GPT-5.4 wins six of the eight areas we test: coding, agents, reasoning, images and charts, math and long documents. GPT-5.4 nano wins none, but costs 92% less. They are level on facts and following instructions.
Which is better for coding?
GPT-5.4. It wins 6 of the 6 coding tests both models report; GPT-5.4 nano wins none.
Which is cheaper?
GPT-5.4 costs $2.50 per million input tokens and $15.00 per million output tokens; GPT-5.4 nano costs $0.20 and $1.25. That makes GPT-5.4 nano about 92% cheaper for the same work.
How do you compare the two?
We use the 45 benchmark tests both models have published scores on. The verdict counts the 29 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.