Model comparison
GPT-5.4 mini vs GPT-5 Nano
GPT-5.4 mini is the stronger model overall, scoring 44.8 to 33.4 on the Noometry Index. GPT-5 Nano costs 12× less per token, which makes it the better buy when GPT-5.4 mini'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 mini scores higher in 9 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 mini leads 64.2 to 39.4.
- The biggest single-benchmark swing is ARC-AGI-1: 63.7% for GPT-5.4 mini and 20.7% for GPT-5 Nano.
- GPT-5 Nano is cheaper at $0.05 / $0.40 per million input/output tokens, against $0.75 / $4.50 for GPT-5.4 mini.
Side by side
| GPT-5.4 mini | GPT-5 Nano | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 44.8 | 33.4 |
| Released | 2026-03-17 | 2025-08-07 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 400K |
| Max output | 128K | 128K |
| Input $ / M tokens | $0.75 | $0.05 |
| Output $ / M tokens | $4.50 | $0.40 |
| Results tracked | 46 | 48 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.4 mini leads
GPT-5.4 mini: 45.0 (#67), GPT-5 Nano: 33.3 (#248)
| Benchmark | GPT-5.4 mini | GPT-5 Nano |
|---|---|---|
| WeirdML | 60.3% | 38.1% |
| LMArena Coding | 1438 | 1351 |
| ALE-Bench | 1,189 | 718.67 |
| FrontierCode | 27% | — |
| SWE-bench Verified (bash only) | — | 34.8% |
| LMArena WebDev | 1397 | — |
| SciCode | 49.9% | — |
Agentic & Tool Use GPT-5.4 mini leads
GPT-5.4 mini: 29.6 (#74), GPT-5 Nano: 25.6 (#99)
| Benchmark | GPT-5.4 mini | GPT-5 Nano |
|---|---|---|
| Terminal-Bench | — | 21.8% |
| Berkeley Function Calling Leaderboard | — | 51.5% |
| DeepResearch Bench | 36.3% | — |
Reasoning GPT-5.4 mini leads
GPT-5.4 mini: 30.3 (#77), GPT-5 Nano: 16.4 (#291)
| Benchmark | GPT-5.4 mini | GPT-5 Nano |
|---|---|---|
| ARC-AGI-2 | 18.9% | 2.6% |
| Kagi LLM Benchmark | 37.9% | 62.2% |
| ARC-AGI-1 | 63.7% | 20.7% |
| Chess Puzzles | 24% | 27% |
| LMArena Hard Prompts | 1424 | 1329 |
| Mystery Game Puzzles | 11% | 9% |
| DTBench | 80% | 62.7% |
| LMCA | 40.8% | 7.9% |
| Epoch Capabilities Index | 148.83 | 139.39 |
| ForecastBench | 57 | 59.1 |
| NYT Connections (extended) | 61.8% | — |
| CritPt | 10% | — |
| Thematic Generalization | 61.7% | — |
Math GPT-5.4 mini leads
GPT-5.4 mini: 45.4 (#63), GPT-5 Nano: 29.1 (#221)
| Benchmark | GPT-5.4 mini | GPT-5 Nano |
|---|---|---|
| FrontierMath (Tiers 1-3) | 51.2% | 20% |
| FrontierMath Tier 4 | 9.8% | 2.4% |
| OTIS Mock AIME 2024-2025 | 88.9% | 81.1% |
| ProofBench | 21% | 12% |
| LMArena Math | 1418 | 1318 |
| FrontierMath (Feb 2025 set) | 28.3% | 8.3% |
| FrontierMath Tier 4 (v1) | 2.1% | 2.1% |
| Omni-MATH | — | 54.6% |
| MATH Level 5 | — | 95.2% |
Knowledge GPT-5.4 mini leads
GPT-5.4 mini: 51.4 (#62), GPT-5 Nano: 35.8 (#166)
| Benchmark | GPT-5.4 mini | GPT-5 Nano |
|---|---|---|
| GPQA Diamond | 86.9% | 69.4% |
| SimpleQA Verified | 29.4% | 11.7% |
| Vectara Hallucination Rate | 5.5% | 10.5% |
| LMArena Expert | 1434 | 1321 |
| MMLU-Pro | — | 77.8% |
| GPQA (HELM) | — | 67.9% |
Multimodal GPT-5.4 mini leads
GPT-5.4 mini: 39.0 (#51), GPT-5 Nano: 30.7 (#102)
| Benchmark | GPT-5.4 mini | GPT-5 Nano |
|---|---|---|
| LMArena Vision | 1245 | 1159 |
| VPCT | — | 37.2% |
Multilingual GPT-5.4 mini leads
GPT-5.4 mini: 52.1 (#94), GPT-5 Nano: 45.5 (#167)
| Benchmark | GPT-5.4 mini | GPT-5 Nano |
|---|---|---|
| LMArena Non-English | 1405 | 1313 |
| LMArena Chinese | 1446 | 1355 |
| LMArena German | 1407 | 1329 |
| LMArena Korean | 1369 | 1269 |
| LMArena Russian | 1417 | 1296 |
| LMArena Spanish | 1403 | 1360 |
| LMArena French | 1440 | — |
| LMArena Japanese | 1375 | — |
Instruction Following Too close to call
GPT-5.4 mini: 74.3 (#99), GPT-5 Nano: 75.2 (#72)
| Benchmark | GPT-5.4 mini | GPT-5 Nano |
|---|---|---|
| LMArena Instruction Following | 1405 | 1306 |
| IFEval | — | 93.2% |
Long Context GPT-5.4 mini leads
GPT-5.4 mini: 42.9 (#107), GPT-5 Nano: 31.3 (#275)
| Benchmark | GPT-5.4 mini | GPT-5 Nano |
|---|---|---|
| LMArena Longer Query | 1407 | 1313 |
| Fiction.LiveBench | — | 44.4% |
Writing & Preference GPT-5.4 mini leads
GPT-5.4 mini: 64.2 (#56), GPT-5 Nano: 39.4 (#244)
| Benchmark | GPT-5.4 mini | GPT-5 Nano |
|---|---|---|
| LMArena Text | 1411 | 1320 |
| LMArena Creative Writing | 1370 | 1250 |
| EQ-Bench Creative Writing | 1665 | 705 |
| LMArena Multi-Turn | 1429 | 1311 |
| WildBench | — | 80.6% |
Frequently asked questions
Is GPT-5.4 mini better than GPT-5 Nano?
GPT-5.4 mini is the stronger model overall, scoring 44.8 to 33.4 on the Noometry Index. GPT-5 Nano costs 12× less per token, which makes it the better buy when GPT-5.4 mini's lead doesn't matter for your workload.
Which is cheaper, GPT-5.4 mini 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 mini lists at $0.75 and $4.50.
Is GPT-5.4 mini or GPT-5 Nano better for coding?
GPT-5.4 mini scores higher on coding benchmarks: 45.0 versus 33.3 in the Noometry coding category.
Which has the bigger context window?
Both accept 400K tokens.
How many benchmarks do GPT-5.4 mini and GPT-5 Nano share?
37 benchmarks have published results for both models. GPT-5.4 mini has 46 scored results on Noometry and GPT-5 Nano has 48.