Model comparison
GPT-4o vs GPT-5 Nano
GPT-5 Nano is the stronger model overall, scoring 33.5 to 28.6 on the Noometry Index.
Last verified . 40 shared benchmarks.
Summary
- They share 40 benchmarks with published results for both. GPT-4o scores higher in 3 categories and GPT-5 Nano in 7 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5 Nano leads 29.4 to 10.6.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 6.4% for GPT-4o and 81.1% for GPT-5 Nano.
- GPT-5 Nano is cheaper at $0.05 / $0.40 per million input/output tokens, against $2.50 / $10 for GPT-4o.
- GPT-5 Nano accepts more context: 400K tokens versus 128K.
Side by side
| GPT-4o | GPT-5 Nano | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 28.6 | 33.5 |
| Released | 2024-05-13 | 2025-08-07 |
| Weights | Proprietary | Proprietary |
| Context window | 128K | 400K |
| Max output | 16K | 128K |
| Input $ / M tokens | $2.50 | $0.05 |
| Output $ / M tokens | $10 | $0.40 |
| Results tracked | 72 | 49 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5 Nano leads
GPT-4o: 24.8 (#328), GPT-5 Nano: 33.6 (#254)
| Benchmark | GPT-4o | GPT-5 Nano |
|---|---|---|
| SWE-bench Verified (bash only) | 21.6% | 34.8% |
| WeirdML | 25.1% | 38.1% |
| LMArena Coding | 1297 | 1351 |
| SWE-bench Verified | 31% | — |
| Aider Polyglot | 45.3% | — |
| GSO | 0% | — |
| BigCodeBench Instruct | 51.1% | — |
| LiveBench Coding | 51.4% | — |
| BigCodeBench Complete | 61.1% | — |
| CadEval | 26% | — |
| ALE-Bench | — | 718.67 |
| HumanEval+ | 87.2% | — |
| MBPP+ | 72.2% | — |
Agentic & Tool Use GPT-5 Nano leads
GPT-4o: 21.0 (#141), GPT-5 Nano: 25.8 (#106)
| Benchmark | GPT-4o | GPT-5 Nano |
|---|---|---|
| Terminal-Bench | — | 21.8% |
| Berkeley Function Calling Leaderboard | — | 51.5% |
| GDPval | 9.9% | — |
| TheAgentCompany | 8.6% | — |
| Cybench | 12.5% | — |
| BALROG | 32.3% | — |
| LMArena Search | 1006 | — |
| METR Time Horizons | 40.8% | — |
Reasoning GPT-5 Nano leads
GPT-4o: 9.4 (#343), GPT-5 Nano: 16.3 (#306)
| Benchmark | GPT-4o | GPT-5 Nano |
|---|---|---|
| ARC-AGI-2 | 0% | 2.6% |
| ARC-AGI-1 | 4.5% | 20.7% |
| Chess Puzzles | 13% | 27% |
| LMArena Hard Prompts | 1281 | 1328 |
| DTBench | 64.5% | 62.7% |
| LMCA | 16.6% | 7.9% |
| Epoch Capabilities Index | 128.97 | 139.38 |
| ForecastBench | 57.7 | 59.1 |
| SimpleBench | 17.8% | — |
| Kagi LLM Benchmark | — | 62.2% |
| CritPt | 0% | — |
| EnigmaEval | 0.8% | — |
| LiveBench Reasoning | 55.8% | — |
| Mystery Game Puzzles | — | 9% |
| LiveBench Data Analysis | 60.9% | — |
| LiveBench | 55.3% | — |
Math GPT-5 Nano leads
GPT-4o: 10.6 (#312), GPT-5 Nano: 29.4 (#241)
| Benchmark | GPT-4o | GPT-5 Nano |
|---|---|---|
| FrontierMath (Tiers 1-3) | 0.4% | 20% |
| OTIS Mock AIME 2024-2025 | 6.4% | 81.1% |
| Omni-MATH | 29.3% | 54.6% |
| LMArena Math | 1285 | 1317 |
| MATH Level 5 | 53.3% | 95.2% |
| FrontierMath (Feb 2025 set) | 0.3% | 8.3% |
| FrontierMath Tier 4 | — | 2.4% |
| ProofBench | — | 12% |
| LiveBench Math | 49.5% | — |
| FrontierMath Tier 4 (v1) | — | 2.1% |
Knowledge GPT-5 Nano leads
GPT-4o: 28.8 (#242), GPT-5 Nano: 35.9 (#178)
| Benchmark | GPT-4o | GPT-5 Nano |
|---|---|---|
| GPQA Diamond | 49.2% | 69.4% |
| SimpleQA Verified | 26% | 11.7% |
| MMLU-Pro | 71.3% | 77.8% |
| Vectara Hallucination Rate | 9.6% | 10.5% |
| GPQA (HELM) | 52% | 67.9% |
| LMArena Expert | 1250 | 1321 |
| Humanity's Last Exam | 2.7% | — |
| Confabulations | 15.3% | — |
| MMLU | 88.1% | — |
Multimodal GPT-4o leads
GPT-4o: 34.5 (#91), GPT-5 Nano: 31.3 (#108)
| Benchmark | GPT-4o | GPT-5 Nano |
|---|---|---|
| LMArena Vision | 1137 | 1159 |
| VPCT | 40% | 37.2% |
| Video-MME | 71.9% | — |
| GeoBench | 71% | — |
| ScienceQA | 88.5% | — |
Multilingual GPT-5 Nano leads
GPT-4o: 43.2 (#186), GPT-5 Nano: 45.3 (#172)
| Benchmark | GPT-4o | GPT-5 Nano |
|---|---|---|
| LMArena Non-English | 1283 | 1313 |
| LMArena Chinese | 1277 | 1356 |
| LMArena German | 1282 | 1327 |
| LMArena Japanese | 1257 | 1226 |
| LMArena Korean | 1234 | 1269 |
| LMArena Russian | 1286 | 1296 |
| LMArena Spanish | 1292 | 1360 |
| LMArena French | 1304 | — |
Instruction Following GPT-5 Nano leads
GPT-4o: 66.6 (#207), GPT-5 Nano: 75.0 (#79)
| Benchmark | GPT-4o | GPT-5 Nano |
|---|---|---|
| IFEval | 81.7% | 93.2% |
| LMArena Instruction Following | 1278 | 1306 |
| LiveBench Instruction Following | 68.6% | — |
Long Context GPT-4o leads
GPT-4o: 39.4 (#179), GPT-5 Nano: 31.3 (#281)
| Benchmark | GPT-4o | GPT-5 Nano |
|---|---|---|
| Fiction.LiveBench | 66.7% | 44.4% |
| LMArena Longer Query | 1289 | 1312 |
Writing & Preference GPT-4o leads
GPT-4o: 52.6 (#166), GPT-5 Nano: 39.1 (#249)
| Benchmark | GPT-4o | GPT-5 Nano |
|---|---|---|
| LMArena Text | 1300 | 1320 |
| LMArena Creative Writing | 1292 | 1249 |
| WildBench | 82.8% | 80.6% |
| LMArena Multi-Turn | 1302 | 1311 |
| Short-Story Creative Writing | 81.8% | — |
| EQ-Bench Creative Writing | — | 705 |
| LiveBench Language | 47.6% | — |
Frequently asked questions
Is GPT-4o better than GPT-5 Nano?
GPT-5 Nano is the stronger model overall, scoring 33.5 to 28.6 on the Noometry Index.
Which is cheaper, GPT-4o or GPT-5 Nano?
GPT-5 Nano is cheaper. It lists at $0.05 per million input tokens and $0.40 per million output tokens; GPT-4o lists at $2.50 and $10.
Is GPT-4o or GPT-5 Nano better for coding?
GPT-5 Nano scores higher on coding benchmarks: 33.6 versus 24.8 in the Noometry coding category.
Which has the bigger context window?
GPT-5 Nano does, with 400K tokens against 128K.
How many benchmarks do GPT-4o and GPT-5 Nano share?
40 benchmarks have published results for both models. GPT-4o has 72 scored results on Noometry and GPT-5 Nano has 49.