Model comparison

GPT-4 vs Step 1o Turbo 202506

Step 1o Turbo 202506 is the stronger model overall, scoring 39.7 to 29.1 on the Noometry Index.

Last verified . 13 shared benchmarks.

GPT-4 OpenAI

29.1

Rank #316 Confirmed

Step 1o Turbo 202506 StepFun

39.7

Rank #160 Confirmed

Summary

  • They share 13 benchmarks with published results for both. GPT-4 scores higher in 0 categories and Step 1o Turbo 202506 in 8 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in math, where Step 1o Turbo 202506 leads 36.6 to 10.8.

Side by side

GPT-4 and Step 1o Turbo 202506 specifications
GPT-4Step 1o Turbo 202506
ProviderOpenAIStepFun
Noometry Index29.139.7
Released2023-03-14—
WeightsProprietaryProprietary
Context window8K—
Max output8K—
Input $ / M tokens$30—
Output $ / M tokens$60—
Results tracked3814

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding Step 1o Turbo 202506 leads

GPT-4: 31.6 (#283), Step 1o Turbo 202506: 39.2 (#160)

Coding benchmarks
BenchmarkGPT-4Step 1o Turbo 202506
LMArena Coding12541339
WeirdML12.4%—
BigCodeBench Instruct46%—
BigCodeBench Complete57.2%—
HumanEval+79.3%—

Agentic & Tool Use Not comparable

GPT-4: —, Step 1o Turbo 202506: —

Agentic & Tool Use benchmarks
BenchmarkGPT-4Step 1o Turbo 202506
METR Time Horizons36.1%—

Reasoning Step 1o Turbo 202506 leads

GPT-4: 17.8 (#289), Step 1o Turbo 202506: 26.8 (#129)

Reasoning benchmarks
BenchmarkGPT-4Step 1o Turbo 202506
LMArena Hard Prompts12411335
Chess Puzzles4%—
Mystery Game Puzzles12%—
DTBench62.7%—
LMCA17.1%—
BIG-Bench Hard75.1%—
Epoch Capabilities Index125.89—
ForecastBench57.8—
HellaSwag95.3%—
WinoGrande87.5%—

Math Step 1o Turbo 202506 leads

GPT-4: 10.8 (#309), Step 1o Turbo 202506: 36.6 (#164)

Math benchmarks
BenchmarkGPT-4Step 1o Turbo 202506
LMArena Math12691318
OTIS Mock AIME 2024-20251.1%—
MATH Level 523%—
GSM8K92%—

Knowledge Step 1o Turbo 202506 leads

GPT-4: 18.4 (#282), Step 1o Turbo 202506: 36.1 (#176)

Knowledge benchmarks
BenchmarkGPT-4Step 1o Turbo 202506
LMArena Expert12111308
GPQA Diamond35.7%—
MMLU86.4%—
TriviaQA84.8%—

Multimodal Not comparable

GPT-4: —, Step 1o Turbo 202506: 36.1 (#80)

Multimodal benchmarks
BenchmarkGPT-4Step 1o Turbo 202506
LMArena Vision—1186

Multilingual Step 1o Turbo 202506 leads

GPT-4: 40.6 (#215), Step 1o Turbo 202506: 45.3 (#173)

Multilingual benchmarks
BenchmarkGPT-4Step 1o Turbo 202506
LMArena Non-English12461313
LMArena Chinese12421380
LMArena German12511314
LMArena Russian12511327
LMArena French1283—
LMArena Japanese1209—
LMArena Korean1184—
LMArena Spanish1261—

Instruction Following Step 1o Turbo 202506 leads

GPT-4: 65.3 (#222), Step 1o Turbo 202506: 69.2 (#176)

Instruction Following benchmarks
BenchmarkGPT-4Step 1o Turbo 202506
LMArena Instruction Following12411310

Long Context Step 1o Turbo 202506 leads

GPT-4: 37.7 (#212), Step 1o Turbo 202506: 41.0 (#146)

Long Context benchmarks
BenchmarkGPT-4Step 1o Turbo 202506
LMArena Longer Query12441348

Writing & Preference Step 1o Turbo 202506 leads

GPT-4: 34.9 (#268), Step 1o Turbo 202506: 53.1 (#160)

Writing & Preference benchmarks
BenchmarkGPT-4Step 1o Turbo 202506
LMArena Text12631336
LMArena Creative Writing12441307
LMArena Multi-Turn12571340
EQ-Bench Creative Writing752—

Frequently asked questions

Is GPT-4 better than Step 1o Turbo 202506?

Step 1o Turbo 202506 is the stronger model overall, scoring 39.7 to 29.1 on the Noometry Index.

Is GPT-4 or Step 1o Turbo 202506 better for coding?

Step 1o Turbo 202506 scores higher on coding benchmarks: 39.2 versus 31.6 in the Noometry coding category.

How many benchmarks do GPT-4 and Step 1o Turbo 202506 share?

13 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and Step 1o Turbo 202506 has 14.

Related comparisons

Go deeper