Model comparison
GPT-5 Nano vs o4-mini
o4-mini is the stronger model overall, scoring 41.6 to 33.5 on the Noometry Index. GPT-5 Nano costs 14× less per token, which makes it the better buy when o4-mini's lead doesn't matter for your workload.
Last verified . 46 shared benchmarks.
Summary
- They share 46 benchmarks with published results for both. GPT-5 Nano scores higher in 0 categories and o4-mini in 10 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where o4-mini leads 54.0 to 39.1.
- The biggest single-benchmark swing is ARC-AGI-1: 20.7% for GPT-5 Nano and 58.7% for o4-mini.
- GPT-5 Nano is cheaper at $0.05 / $0.40 per million input/output tokens, against $1.10 / $4.40 for o4-mini.
- GPT-5 Nano accepts more context: 400K tokens versus 200K.
Side by side
| GPT-5 Nano | o4-mini | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 33.5 | 41.6 |
| Released | 2025-08-07 | 2025-04-16 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 200K |
| Max output | 128K | 100K |
| Input $ / M tokens | $0.05 | $1.10 |
| Output $ / M tokens | $0.40 | $4.40 |
| Results tracked | 49 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o4-mini leads
GPT-5 Nano: 33.6 (#254), o4-mini: 40.9 (#127)
| Benchmark | GPT-5 Nano | o4-mini |
|---|---|---|
| SWE-bench Verified (bash only) | 34.8% | 45% |
| WeirdML | 38.1% | 52.6% |
| LMArena Coding | 1351 | 1368 |
| ALE-Bench | 718.67 | 826.17 |
| Aider Polyglot | — | 72% |
| GSO | — | 3.6% |
| CadEval | — | 62% |
| AlgoTune | — | 1.72 |
Agentic & Tool Use o4-mini leads
GPT-5 Nano: 25.8 (#106), o4-mini: 32.6 (#61)
| Benchmark | GPT-5 Nano | o4-mini |
|---|---|---|
| Berkeley Function Calling Leaderboard | 51.5% | 53.2% |
| Terminal-Bench | 21.8% | — |
| GDPval | — | 25.3% |
| METR Time Horizons | — | 63.9% |
Reasoning o4-mini leads
GPT-5 Nano: 16.3 (#306), o4-mini: 24.6 (#162)
| Benchmark | GPT-5 Nano | o4-mini |
|---|---|---|
| ARC-AGI-2 | 2.6% | 6.1% |
| Kagi LLM Benchmark | 62.2% | 67.6% |
| ARC-AGI-1 | 20.7% | 58.7% |
| Chess Puzzles | 27% | 26% |
| LMArena Hard Prompts | 1328 | 1351 |
| Mystery Game Puzzles | 9% | 5% |
| DTBench | 62.7% | 77.6% |
| LMCA | 7.9% | 26.5% |
| Epoch Capabilities Index | 139.38 | 145.64 |
| ForecastBench | 59.1 | 61.8 |
| SimpleBench | — | 38.7% |
| CritPt | — | 0.6% |
| EnigmaEval | — | 9.2% |
Math o4-mini leads
GPT-5 Nano: 29.4 (#241), o4-mini: 40.8 (#89)
| Benchmark | GPT-5 Nano | o4-mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 20% | 36.1% |
| FrontierMath Tier 4 | 2.4% | 4.9% |
| OTIS Mock AIME 2024-2025 | 81.1% | 81.7% |
| Omni-MATH | 54.6% | 72% |
| LMArena Math | 1317 | 1389 |
| MATH Level 5 | 95.2% | 97.8% |
| FrontierMath (Feb 2025 set) | 8.3% | 24.8% |
| FrontierMath Tier 4 (v1) | 2.1% | 6.3% |
| ProofBench | 12% | — |
Knowledge o4-mini leads
GPT-5 Nano: 35.9 (#178), o4-mini: 43.6 (#91)
| Benchmark | GPT-5 Nano | o4-mini |
|---|---|---|
| GPQA Diamond | 69.4% | 79.6% |
| SimpleQA Verified | 11.7% | 19.6% |
| MMLU-Pro | 77.8% | 82% |
| Vectara Hallucination Rate | 10.5% | 18.6% |
| GPQA (HELM) | 67.9% | 73.5% |
| LMArena Expert | 1321 | 1343 |
| Humanity's Last Exam | — | 18.1% |
| Confabulations | — | 15.8% |
Multimodal o4-mini leads
GPT-5 Nano: 31.3 (#108), o4-mini: 40.2 (#49)
| Benchmark | GPT-5 Nano | o4-mini |
|---|---|---|
| LMArena Vision | 1159 | 1194 |
| VPCT | 37.2% | 57.5% |
| GeoBench | — | 64% |
Multilingual o4-mini leads
GPT-5 Nano: 45.3 (#172), o4-mini: 47.0 (#154)
| Benchmark | GPT-5 Nano | o4-mini |
|---|---|---|
| LMArena Non-English | 1313 | 1337 |
| LMArena Chinese | 1356 | 1354 |
| LMArena German | 1327 | 1336 |
| LMArena Japanese | 1226 | 1308 |
| LMArena Korean | 1269 | 1312 |
| LMArena Russian | 1296 | 1334 |
| LMArena Spanish | 1360 | 1347 |
| LMArena French | — | 1364 |
Instruction Following Too close to call
GPT-5 Nano: 75.0 (#79), o4-mini: 75.2 (#68)
| Benchmark | GPT-5 Nano | o4-mini |
|---|---|---|
| IFEval | 93.2% | 92.8% |
| LMArena Instruction Following | 1306 | 1321 |
Long Context o4-mini leads
GPT-5 Nano: 31.3 (#281), o4-mini: 45.5 (#33)
| Benchmark | GPT-5 Nano | o4-mini |
|---|---|---|
| Fiction.LiveBench | 44.4% | 77.8% |
| LMArena Longer Query | 1312 | 1315 |
Writing & Preference o4-mini leads
GPT-5 Nano: 39.1 (#249), o4-mini: 54.0 (#152)
| Benchmark | GPT-5 Nano | o4-mini |
|---|---|---|
| LMArena Text | 1320 | 1353 |
| LMArena Creative Writing | 1249 | 1294 |
| WildBench | 80.6% | 85.4% |
| LMArena Multi-Turn | 1311 | 1350 |
| Short-Story Creative Writing | — | 75% |
| EQ-Bench Creative Writing | 705 | — |
Frequently asked questions
Is GPT-5 Nano better than o4-mini?
o4-mini is the stronger model overall, scoring 41.6 to 33.5 on the Noometry Index. GPT-5 Nano costs 14× less per token, which makes it the better buy when o4-mini's lead doesn't matter for your workload.
Which is cheaper, GPT-5 Nano or o4-mini?
GPT-5 Nano is cheaper. It lists at $0.05 per million input tokens and $0.40 per million output tokens; o4-mini lists at $1.10 and $4.40.
Is GPT-5 Nano or o4-mini better for coding?
o4-mini scores higher on coding benchmarks: 40.9 versus 33.6 in the Noometry coding category.
Which has the bigger context window?
GPT-5 Nano does, with 400K tokens against 200K.
How many benchmarks do GPT-5 Nano and o4-mini share?
46 benchmarks have published results for both models. GPT-5 Nano has 49 scored results on Noometry and o4-mini has 60.