Learn about the capabilities of different GPT-3 engines.

We'll cover the following

Common GPT-3 engines

OpenAI’s GPT-3 model has many engines, and every engine has different capabilities. The most commonly used engines of GPT-3 are the following:

Engine

Strengths

Rank

Description

text-davinci-002

Creative generation, search, cause and effect, summarization, complex intent

Most Capable/

Slowest

It can perform any task that other GPT-3 models can with less context. It can take maximum tokens up to 4000 per request.

text-curie-001

Language translation, complex classification, sentiment, summarization

Very Capable/

Faster

It is very capable of the two given below and has a lower cost than text-davinci-002. It can take maximum tokens up to 2048 per request.

text-babbage-001

Moderate classification, semantic search

Capable for Straightforward tasks/

Very Fast

Lower cost model to perform straightforward tasks and it is faster than the above two. It can take maximum tokens up to 2048 per request.

text-ada-001

Parsing text, simple classification, address correction, keywords

Capable of Simple tasks/

Fastest

It is the fastest model of all GPT-3 models and has the lowest cost. It can take maximum tokens up to 2048 per request.

Get hands-on with 1200+ tech skills courses.