# GPT-4 vs Step 5 Preview

> Step 5 Preview is the stronger model overall, scoring 47.9 to 29.1 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gpt-4-vs-step-5-preview
- Last updated: 2026-10-10
- Shared benchmarks: 13

## Summary

- They share 13 benchmarks with published results for both. GPT-4 scores higher in 0 categories and Step 5 Preview in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Step 5 Preview leads 46.1 to 10.8.
- Step 5 Preview is cheaper at $1 / $2.70 per million input/output tokens, against $30 / $60 for GPT-4.
- Step 5 Preview accepts more context: 1.02M tokens versus 8K.

## Snapshot

| | GPT-4 | Step 5 Preview |
|---|---|---|
| Provider | OpenAI | StepFun |
| Noometry Index | 29.1 | 47.9 |
| Rank | 316 | 58 |
| Context | 8K | 1.02M |
| Input $/M | $30 | $1 |
| Output $/M | $60 | $2.70 |
| Weights | Proprietary | Proprietary |

## Coding

- GPT-4: 31.6 (#283)
- Step 5 Preview: 51.8 (#36)

| Benchmark | GPT-4 | Step 5 Preview |
|---|---|---|
| LMArena Coding | 1254 | 1480 |
| LMArena WebDev | — | 1564 |
| SciCode | — | 58.9% |
| WeirdML | 12.4% | — |
| BigCodeBench Instruct | 46% | — |
| BigCodeBench Complete | 57.2% | — |
| HumanEval+ | 79.3% | — |

## Agentic & Tool Use

- GPT-4: —
- Step 5 Preview: —

| Benchmark | GPT-4 | Step 5 Preview |
|---|---|---|
| METR Time Horizons | 36.1% | — |

## Reasoning

- GPT-4: 17.8 (#289)
- Step 5 Preview: 40.0 (#57)

| Benchmark | GPT-4 | Step 5 Preview |
|---|---|---|
| LMArena Hard Prompts | 1241 | 1465 |
| CritPt | — | 20.9% |
| Chess Puzzles | 4% | — |
| Mystery Game Puzzles | 12% | — |
| DTBench | 62.7% | — |
| LMCA | 17.1% | — |
| BIG-Bench Hard | 75.1% | — |
| Epoch Capabilities Index | 125.89 | — |
| ForecastBench | 57.8 | — |
| HellaSwag | 95.3% | — |
| WinoGrande | 87.5% | — |

## Math

- GPT-4: 10.8 (#309)
- Step 5 Preview: 46.1 (#72)

| Benchmark | GPT-4 | Step 5 Preview |
|---|---|---|
| LMArena Math | 1269 | 1470 |
| OTIS Mock AIME 2024-2025 | 1.1% | — |
| ProofBench | — | 42% |
| MATH Level 5 | 23% | — |
| GSM8K | 92% | — |

## Knowledge

- GPT-4: 18.4 (#282)
- Step 5 Preview: 41.2 (#112)

| Benchmark | GPT-4 | Step 5 Preview |
|---|---|---|
| LMArena Expert | 1211 | 1470 |
| GPQA Diamond | 35.7% | — |
| MMLU | 86.4% | — |
| TriviaQA | 84.8% | — |

## Multimodal

- GPT-4: —
- Step 5 Preview: 41.0 (#41)

| Benchmark | GPT-4 | Step 5 Preview |
|---|---|---|
| LMArena Vision | — | 1267 |

## Multilingual

- GPT-4: 40.6 (#215)
- Step 5 Preview: 53.8 (#54)

| Benchmark | GPT-4 | Step 5 Preview |
|---|---|---|
| LMArena Non-English | 1246 | 1432 |
| LMArena Chinese | 1242 | 1519 |
| LMArena Russian | 1251 | 1436 |
| LMArena Spanish | 1261 | 1439 |
| LMArena French | 1283 | — |
| LMArena German | 1251 | — |
| LMArena Japanese | 1209 | — |
| LMArena Korean | 1184 | — |

## Instruction Following

- GPT-4: 65.3 (#222)
- Step 5 Preview: 76.0 (#50)

| Benchmark | GPT-4 | Step 5 Preview |
|---|---|---|
| LMArena Instruction Following | 1241 | 1444 |

## Long Context

- GPT-4: 37.7 (#212)
- Step 5 Preview: 44.5 (#64)

| Benchmark | GPT-4 | Step 5 Preview |
|---|---|---|
| LMArena Longer Query | 1244 | 1455 |

## Writing & Preference

- GPT-4: 34.9 (#268)
- Step 5 Preview: 62.8 (#71)

| Benchmark | GPT-4 | Step 5 Preview |
|---|---|---|
| LMArena Text | 1263 | 1442 |
| LMArena Creative Writing | 1244 | 1410 |
| LMArena Multi-Turn | 1257 | 1447 |
| EQ-Bench Creative Writing | 752 | — |

## FAQ

### Is GPT-4 better than Step 5 Preview?

Step 5 Preview is the stronger model overall, scoring 47.9 to 29.1 on the Noometry Index.

### Which is cheaper, GPT-4 or Step 5 Preview?

Step 5 Preview is cheaper. It lists at $1 per million input tokens and $2.70 per million output tokens; GPT-4 lists at $30 and $60.

### Is GPT-4 or Step 5 Preview better for coding?

Step 5 Preview scores higher on coding benchmarks: 51.8 versus 31.6 in the Noometry coding category.

### Which has the bigger context window?

Step 5 Preview does, with 1.02M tokens against 8K.

### How many benchmarks do GPT-4 and Step 5 Preview share?

13 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and Step 5 Preview has 18.