Enumeration OpenAiModel

Capable of very simple tasks, usually the fastest model in the GPT-3 series, and lowest cost.

String presentation recognized by the OpenAI REST API: text-ada-001

Recommended default values:

• max_tokens: 2,048
• temperature: 0.5
• top_p: 1
• frequency_penalty: 0
• presence_penalty: 0

Link for extra information: https://beta.openai.com/docs/models/gpt-3

Remarks:

• The max_tokens parameter specifies the maximum number of tokens that the model is allowed to generate in its response.

Alias names: ada (older version)

Capable of straightforward tasks, very fast, and lower cost.

String presentation recognized by the OpenAI REST API: text-babbage-001

Recommended default values:

• max_tokens: 2,048
• temperature: 0.5
• top_p: 1
• frequency_penalty: 0
• presence_penalty: 0

Link for extra information: https://beta.openai.com/docs/models/gpt-3

Remarks:

• The max_tokens parameter specifies the maximum number of tokens that the model is allowed to generate in its response.

Alias names: babbage (older version)

A set of models that can understand and generate code, including translating natural language to code (Limited Beta).

String presentation recognized by the OpenAI REST API: codex

Link for extra information: https://beta.openai.com/docs/models/codex

Recommended default values:

• max_tokens: 8000
• temperature: 0.5
• top_p: 1
• frequency_penalty: 0
• presence_penalty: 0

Remarks:

• This model is currently in limited beta.

A fine-tuned model that can detect whether text may be sensitive or unsafe.

String presentation recognized by the OpenAI REST API: content-filter

Recommended default values:

• max_tokens: 1
• temperature: 0.0
• top_p: 0
• frequency_penalty: 0
• presence_penalty: 0
• logprobs: 0

Wrap your prompt in the following way:

"<|endoftext|>[prompt]\n--\nLabel:"

Link for extra information: https://beta.openai.com/docs/models/content-filter

Very capable GPT-3 model, but faster and lower cost than Davinci.

String presentation recognized by the OpenAI REST API: text-curie-001

Recommended default values:

• max_tokens: 2,048
• temperature: 0.5
• top_p: 1
• frequency_penalty: 0
• presence_penalty: 0

Link for extra information: https://beta.openai.com/docs/models/gpt-3

Remarks:

• This model can perform many of the same tasks as Davinci, but faster and for 1/10th the cost.
• The max_tokens parameter specifies the maximum number of tokens that the model is allowed to generate in its response.

Alias names: curie (older version)

The most capable GPT-3 model. Can do any task the other models can do, often with higher quality, longer output and better instruction-following. Also supports inserting completions within text.

String presentation recognized by the OpenAI REST API: text-davinci-003

Recommended default values:

• max_tokens: 4,000
• temperature: 0.5
• top_p: 1
• frequency_penalty: 0
• presence_penalty: 0

Link for extra information: https://beta.openai.com/docs/models/gpt-3

Remarks:

• This model is generally the most capable in the GPT-3 series.
• The max_tokens parameter specifies the maximum number of tokens that the model is allowed to generate in its response.

Alias names: davinci (older version)