# GPT-5.4 nano vs Step 3.7 Flash

> GPT-5.4 nano is the stronger model overall, scoring 41.9 to 36.5 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gpt-5-4-nano-vs-step-3-7-flash
- Last updated: 2026-10-10
- Shared benchmarks: 3

## Summary

- They share 3 benchmarks with published results for both. GPT-5.4 nano scores higher in 2 categories and Step 3.7 Flash in 1 category; 2 gaps are clear of the uncertainty.
- The widest gap is in coding, where GPT-5.4 nano leads 43.1 to 39.4.
- The biggest single-benchmark swing is CritPt: 9.3% for GPT-5.4 nano and 2.3% for Step 3.7 Flash.
- Step 3.7 Flash is cheaper at $0.18 / $1.11 per million input/output tokens, against $0.20 / $1.25 for GPT-5.4 nano.
- GPT-5.4 nano accepts more context: 400K tokens versus 256K.
- Step 3.7 Flash has downloadable open weights; the other is API-only.

## Snapshot

| | GPT-5.4 nano | Step 3.7 Flash |
|---|---|---|
| Provider | OpenAI | StepFun |
| Noometry Index | 41.9 | 36.5 |
| Rank | 111 | 207 |
| Context | 400K | 256K |
| Input $/M | $0.20 | $0.18 |
| Output $/M | $1.25 | $1.11 |
| Weights | Proprietary | Open |

## Coding

- GPT-5.4 nano: 43.1 (#82)
- Step 3.7 Flash: 39.4 (#147)

| Benchmark | GPT-5.4 nano | Step 3.7 Flash |
|---|---|---|
| SciCode | 46.9% | 40% |
| ALE-Bench | 1,005 | 694.12 |
| WeirdML | 49.2% | — |
| LMArena Coding | 1404 | — |

## Reasoning

- GPT-5.4 nano: 23.8 (#159)
- Step 3.7 Flash: 20.9 (#211)

| Benchmark | GPT-5.4 nano | Step 3.7 Flash |
|---|---|---|
| CritPt | 9.3% | 2.3% |
| ARC-AGI-2 | 5.7% | — |
| Kagi LLM Benchmark | 39.7% | — |
| NYT Connections (extended) | — | 39.7% |
| ARC-AGI-1 | 51.5% | — |
| Chess Puzzles | 30% | — |
| LMArena Hard Prompts | 1380 | — |
| Mystery Game Puzzles | 9% | — |
| DTBench | 80.3% | — |
| LMCA | 36.9% | — |
| Epoch Capabilities Index | 145.78 | — |
| ForecastBench | 57.3 | — |

## Math

- GPT-5.4 nano: 40.7 (#76)
- Step 3.7 Flash: 40.9 (#74)

| Benchmark | GPT-5.4 nano | Step 3.7 Flash |
|---|---|---|
| FrontierMath (Tiers 1-3) | 44.9% | — |
| FrontierMath Tier 4 | 12.2% | — |
| MathArena Final-Answer Competitions | — | 68.5% |
| OTIS Mock AIME 2024-2025 | 87.8% | — |
| ProofBench | 5% | — |
| LMArena Math | 1406 | — |
| FrontierMath (Feb 2025 set) | 25.9% | — |
| FrontierMath Tier 4 (v1) | 6.3% | — |

## Knowledge

- GPT-5.4 nano: 41.8 (#94)
- Step 3.7 Flash: —

| Benchmark | GPT-5.4 nano | Step 3.7 Flash |
|---|---|---|
| GPQA Diamond | 78.5% | — |
| SimpleQA Verified | 11.7% | — |
| Vectara Hallucination Rate | 3.1% | — |
| LMArena Expert | 1397 | — |

## Multimodal

- GPT-5.4 nano: 36.0 (#73)
- Step 3.7 Flash: —

| Benchmark | GPT-5.4 nano | Step 3.7 Flash |
|---|---|---|
| LMArena Vision | 1196 | — |

## Multilingual

- GPT-5.4 nano: 48.8 (#136)
- Step 3.7 Flash: —

| Benchmark | GPT-5.4 nano | Step 3.7 Flash |
|---|---|---|
| LMArena Non-English | 1359 | — |
| LMArena Chinese | 1391 | — |
| LMArena French | 1396 | — |
| LMArena German | 1368 | — |
| LMArena Japanese | 1342 | — |
| LMArena Korean | 1321 | — |
| LMArena Russian | 1363 | — |
| LMArena Spanish | 1368 | — |

## Instruction Following

- GPT-5.4 nano: 72.1 (#139)
- Step 3.7 Flash: —

| Benchmark | GPT-5.4 nano | Step 3.7 Flash |
|---|---|---|
| LMArena Instruction Following | 1362 | — |

## Long Context

- GPT-5.4 nano: 41.6 (#131)
- Step 3.7 Flash: —

| Benchmark | GPT-5.4 nano | Step 3.7 Flash |
|---|---|---|
| LMArena Longer Query | 1365 | — |

## Writing & Preference

- GPT-5.4 nano: 55.9 (#137)
- Step 3.7 Flash: —

| Benchmark | GPT-5.4 nano | Step 3.7 Flash |
|---|---|---|
| LMArena Text | 1372 | — |
| LMArena Creative Writing | 1313 | — |
| LMArena Multi-Turn | 1382 | — |

## FAQ

### Is GPT-5.4 nano better than Step 3.7 Flash?

GPT-5.4 nano is the stronger model overall, scoring 41.9 to 36.5 on the Noometry Index.

### Which is cheaper, GPT-5.4 nano or Step 3.7 Flash?

Step 3.7 Flash is cheaper. It lists at $0.18 per million input tokens and $1.11 per million output tokens; GPT-5.4 nano lists at $0.20 and $1.25.

### Is GPT-5.4 nano or Step 3.7 Flash better for coding?

GPT-5.4 nano scores higher on coding benchmarks: 43.1 versus 39.4 in the Noometry coding category.

### Which has the bigger context window?

GPT-5.4 nano does, with 400K tokens against 256K.

### How many benchmarks do GPT-5.4 nano and Step 3.7 Flash share?

3 benchmarks have published results for both models. GPT-5.4 nano has 40 scored results on Noometry and Step 3.7 Flash has 5.