# GPT-4 vs GPT-4.1 nano

> GPT-4 is the stronger model overall, scoring 29.1 to 27.9 on the Noometry Index. GPT-4.1 nano costs 214× less per token, which makes it the better buy when GPT-4's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/gpt-4-vs-gpt-4-1-nano
- Last updated: 2026-10-12
- Shared benchmarks: 22

## Summary

- They share 22 benchmarks with published results for both. GPT-4 scores higher in 3 categories and GPT-4.1 nano in 5 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-4.1 nano leads 26.9 to 10.8.
- The biggest single-benchmark swing is MATH Level 5: 23% for GPT-4 and 70% for GPT-4.1 nano.
- GPT-4.1 nano is cheaper at $0.10 / $0.40 per million input/output tokens, against $30 / $60 for GPT-4.
- GPT-4.1 nano accepts more context: 1.05M tokens versus 8K.

## Snapshot

| | GPT-4 | GPT-4.1 nano |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 29.1 | 27.9 |
| Rank | 316 | 327 |
| Context | 8K | 1.05M |
| Input $/M | $30 | $0.10 |
| Output $/M | $60 | $0.40 |
| Weights | Proprietary | Proprietary |

## Coding

- GPT-4: 31.6 (#283)
- GPT-4.1 nano: 24.1 (#330)

| Benchmark | GPT-4 | GPT-4.1 nano |
|---|---|---|
| WeirdML | 12.4% | 19% |
| LMArena Coding | 1254 | 1306 |
| Aider Polyglot | — | 8.9% |
| SciCode | — | 25.9% |
| BigCodeBench Instruct | 46% | — |
| BigCodeBench Complete | 57.2% | — |
| HumanEval+ | 79.3% | — |

## Agentic & Tool Use

- GPT-4: —
- GPT-4.1 nano: 26.5 (#104)

| Benchmark | GPT-4 | GPT-4.1 nano |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 33% |
| METR Time Horizons | 36.1% | — |

## Reasoning

- GPT-4: 17.8 (#289)
- GPT-4.1 nano: 8.5 (#349)

| Benchmark | GPT-4 | GPT-4.1 nano |
|---|---|---|
| LMArena Hard Prompts | 1241 | 1286 |
| DTBench | 62.7% | 52.5% |
| LMCA | 17.1% | 5.5% |
| Epoch Capabilities Index | 125.89 | 129.62 |
| ARC-AGI-2 | — | 0% |
| Kagi LLM Benchmark | — | 33.3% |
| ARC-AGI-1 | — | 0% |
| CritPt | — | 0% |
| Chess Puzzles | 4% | — |
| Mystery Game Puzzles | 12% | — |
| BIG-Bench Hard | 75.1% | — |
| ForecastBench | 57.8 | — |
| HellaSwag | 95.3% | — |
| WinoGrande | 87.5% | — |

## Math

- GPT-4: 10.8 (#309)
- GPT-4.1 nano: 26.9 (#252)

| Benchmark | GPT-4 | GPT-4.1 nano |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.1% | 28.9% |
| LMArena Math | 1269 | 1274 |
| MATH Level 5 | 23% | 70% |
| Omni-MATH | — | 36.7% |
| FrontierMath (Feb 2025 set) | — | 1% |
| GSM8K | 92% | — |

## Knowledge

- GPT-4: 18.4 (#282)
- GPT-4.1 nano: 21.8 (#273)

| Benchmark | GPT-4 | GPT-4.1 nano |
|---|---|---|
| GPQA Diamond | 35.7% | 48.9% |
| LMArena Expert | 1211 | 1272 |
| SimpleQA Verified | — | 6% |
| MMLU-Pro | — | 55% |
| GPQA (HELM) | — | 50.7% |
| MMLU | 86.4% | — |
| TriviaQA | 84.8% | — |

## Multimodal

- GPT-4: —
- GPT-4.1 nano: 29.2 (#113)

| Benchmark | GPT-4 | GPT-4.1 nano |
|---|---|---|
| LMArena Vision | — | 1063 |

## Multilingual

- GPT-4: 40.6 (#215)
- GPT-4.1 nano: 41.6 (#205)

| Benchmark | GPT-4 | GPT-4.1 nano |
|---|---|---|
| LMArena Non-English | 1246 | 1260 |
| LMArena Chinese | 1242 | 1270 |
| LMArena German | 1251 | 1288 |
| LMArena Japanese | 1209 | 1198 |
| LMArena Russian | 1251 | 1261 |
| LMArena French | 1283 | — |
| LMArena Korean | 1184 | — |
| LMArena Spanish | 1261 | — |

## Instruction Following

- GPT-4: 65.3 (#222)
- GPT-4.1 nano: 67.8 (#193)

| Benchmark | GPT-4 | GPT-4.1 nano |
|---|---|---|
| LMArena Instruction Following | 1241 | 1267 |
| IFEval | — | 84.3% |

## Long Context

- GPT-4: 37.7 (#212)
- GPT-4.1 nano: 23.7 (#296)

| Benchmark | GPT-4 | GPT-4.1 nano |
|---|---|---|
| LMArena Longer Query | 1244 | 1283 |
| Fiction.LiveBench | — | 25% |

## Writing & Preference

- GPT-4: 34.9 (#268)
- GPT-4.1 nano: 40.5 (#243)

| Benchmark | GPT-4 | GPT-4.1 nano |
|---|---|---|
| LMArena Text | 1263 | 1285 |
| LMArena Creative Writing | 1244 | 1260 |
| EQ-Bench Creative Writing | 752 | 946 |
| LMArena Multi-Turn | 1257 | 1277 |
| WildBench | — | 81.2% |

## FAQ

### Is GPT-4 better than GPT-4.1 nano?

GPT-4 is the stronger model overall, scoring 29.1 to 27.9 on the Noometry Index. GPT-4.1 nano costs 214× less per token, which makes it the better buy when GPT-4's lead doesn't matter for your workload.

### Which is cheaper, GPT-4 or GPT-4.1 nano?

GPT-4.1 nano is cheaper. It lists at $0.10 per million input tokens and $0.40 per million output tokens; GPT-4 lists at $30 and $60.

### Is GPT-4 or GPT-4.1 nano better for coding?

GPT-4 scores higher on coding benchmarks: 31.6 versus 24.1 in the Noometry coding category.

### Which has the bigger context window?

GPT-4.1 nano does, with 1.05M tokens against 8K.

### How many benchmarks do GPT-4 and GPT-4.1 nano share?

22 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and GPT-4.1 nano has 38.