Model comparison
GPT-5.4 nano vs GPT-5 Nano
GPT-5.4 nano is the stronger model overall, scoring 41.9 to 33.5 on the Noometry Index. GPT-5 Nano costs 3.4× less per token, which makes it the better buy when GPT-5.4 nano's lead doesn't matter for your workload.
Last verified . 37 shared benchmarks.
Summary
- They share 37 benchmarks with published results for both. GPT-5.4 nano scores higher in 8 categories and GPT-5 Nano in 1 category; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where GPT-5.4 nano leads 55.7 to 39.1.
- The biggest single-benchmark swing is ARC-AGI-1: 51.5% for GPT-5.4 nano and 20.7% for GPT-5 Nano.
- GPT-5 Nano is cheaper at $0.05 / $0.40 per million input/output tokens, against $0.20 / $1.25 for GPT-5.4 nano.
Side by side
| GPT-5.4 nano | GPT-5 Nano | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 41.9 | 33.5 |
| Released | 2026-03-17 | 2025-08-07 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 400K |
| Max output | 128K | 128K |
| Input $ / M tokens | $0.20 | $0.05 |
| Output $ / M tokens | $1.25 | $0.40 |
| Results tracked | 40 | 49 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.4 nano leads
GPT-5.4 nano: 43.6 (#84), GPT-5 Nano: 33.6 (#254)
| Benchmark | GPT-5.4 nano | GPT-5 Nano |
|---|---|---|
| WeirdML | 49.2% | 38.1% |
| LMArena Coding | 1405 | 1351 |
| ALE-Bench | 1,005 | 718.67 |
| SWE-bench Verified (bash only) | — | 34.8% |
| SciCode | 46.9% | — |
Agentic & Tool Use Not comparable
GPT-5.4 nano: —, GPT-5 Nano: 25.8 (#106)
| Benchmark | GPT-5.4 nano | GPT-5 Nano |
|---|---|---|
| Terminal-Bench | — | 21.8% |
| Berkeley Function Calling Leaderboard | — | 51.5% |
Reasoning GPT-5.4 nano leads
GPT-5.4 nano: 23.7 (#173), GPT-5 Nano: 16.3 (#306)
| Benchmark | GPT-5.4 nano | GPT-5 Nano |
|---|---|---|
| ARC-AGI-2 | 5.7% | 2.6% |
| Kagi LLM Benchmark | 39.7% | 62.2% |
| ARC-AGI-1 | 51.5% | 20.7% |
| Chess Puzzles | 30% | 27% |
| LMArena Hard Prompts | 1381 | 1328 |
| Mystery Game Puzzles | 9% | 9% |
| DTBench | 80.3% | 62.7% |
| LMCA | 36.9% | 7.9% |
| Epoch Capabilities Index | 145.81 | 139.38 |
| ForecastBench | 57.3 | 59.1 |
| CritPt | 9.3% | — |
Math GPT-5.4 nano leads
GPT-5.4 nano: 40.9 (#88), GPT-5 Nano: 29.4 (#241)
| Benchmark | GPT-5.4 nano | GPT-5 Nano |
|---|---|---|
| FrontierMath (Tiers 1-3) | 44.9% | 20% |
| FrontierMath Tier 4 | 12.2% | 2.4% |
| OTIS Mock AIME 2024-2025 | 87.8% | 81.1% |
| ProofBench | 5% | 12% |
| LMArena Math | 1406 | 1317 |
| FrontierMath (Feb 2025 set) | 25.9% | 8.3% |
| FrontierMath Tier 4 (v1) | 6.3% | 2.1% |
| Omni-MATH | — | 54.6% |
| MATH Level 5 | — | 95.2% |
Knowledge GPT-5.4 nano leads
GPT-5.4 nano: 41.9 (#103), GPT-5 Nano: 35.9 (#178)
| Benchmark | GPT-5.4 nano | GPT-5 Nano |
|---|---|---|
| GPQA Diamond | 78.5% | 69.4% |
| SimpleQA Verified | 11.7% | 11.7% |
| Vectara Hallucination Rate | 3.1% | 10.5% |
| LMArena Expert | 1396 | 1321 |
| MMLU-Pro | — | 77.8% |
| GPQA (HELM) | — | 67.9% |
Multimodal GPT-5.4 nano leads
GPT-5.4 nano: 36.7 (#78), GPT-5 Nano: 31.3 (#108)
| Benchmark | GPT-5.4 nano | GPT-5 Nano |
|---|---|---|
| LMArena Vision | 1196 | 1159 |
| VPCT | — | 37.2% |
Multilingual GPT-5.4 nano leads
GPT-5.4 nano: 48.6 (#140), GPT-5 Nano: 45.3 (#172)
| Benchmark | GPT-5.4 nano | GPT-5 Nano |
|---|---|---|
| LMArena Non-English | 1359 | 1313 |
| LMArena Chinese | 1392 | 1356 |
| LMArena German | 1367 | 1327 |
| LMArena Japanese | 1343 | 1226 |
| LMArena Korean | 1320 | 1269 |
| LMArena Russian | 1363 | 1296 |
| LMArena Spanish | 1371 | 1360 |
| LMArena French | 1396 | — |
Instruction Following GPT-5 Nano leads
GPT-5.4 nano: 71.9 (#144), GPT-5 Nano: 75.0 (#79)
| Benchmark | GPT-5.4 nano | GPT-5 Nano |
|---|---|---|
| LMArena Instruction Following | 1362 | 1306 |
| IFEval | — | 93.2% |
Long Context GPT-5.4 nano leads
GPT-5.4 nano: 41.6 (#137), GPT-5 Nano: 31.3 (#281)
| Benchmark | GPT-5.4 nano | GPT-5 Nano |
|---|---|---|
| LMArena Longer Query | 1366 | 1312 |
| Fiction.LiveBench | — | 44.4% |
Writing & Preference GPT-5.4 nano leads
GPT-5.4 nano: 55.7 (#142), GPT-5 Nano: 39.1 (#249)
| Benchmark | GPT-5.4 nano | GPT-5 Nano |
|---|---|---|
| LMArena Text | 1372 | 1320 |
| LMArena Creative Writing | 1314 | 1249 |
| LMArena Multi-Turn | 1382 | 1311 |
| EQ-Bench Creative Writing | — | 705 |
| WildBench | — | 80.6% |
Frequently asked questions
Is GPT-5.4 nano better than GPT-5 Nano?
GPT-5.4 nano is the stronger model overall, scoring 41.9 to 33.5 on the Noometry Index. GPT-5 Nano costs 3.4× less per token, which makes it the better buy when GPT-5.4 nano's lead doesn't matter for your workload.
Which is cheaper, GPT-5.4 nano 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-5.4 nano lists at $0.20 and $1.25.
Is GPT-5.4 nano or GPT-5 Nano better for coding?
GPT-5.4 nano scores higher on coding benchmarks: 43.6 versus 33.6 in the Noometry coding category.
Which has the bigger context window?
Both accept 400K tokens.
How many benchmarks do GPT-5.4 nano and GPT-5 Nano share?
37 benchmarks have published results for both models. GPT-5.4 nano has 40 scored results on Noometry and GPT-5 Nano has 49.