Implementation:Cohere ai Cohere python Generation Model

Knowledge Sources	Cohere Python SDK
Domains	SDK, Text Generation
Last Updated	2026-02-15 14:00 GMT

Overview

Generation is a Pydantic model representing the non-streaming response from the Cohere Generate API, containing the generated text results and associated metadata.

Description

The Generation class encapsulates the complete response from the Cohere Generate API when called without streaming. It contains:

• id: A unique identifier for the generation request
• prompt: The original prompt text used for generation (optional, may not always be returned)
• generations: A list of SingleGeneration objects, where each entry represents one generated result. When num_generations is greater than 1, this list contains multiple results. Each SingleGeneration includes its own id, text, optional index, optional likelihood, and optional token_likelihoods.
• meta: Optional API metadata including token counts and warnings

The class extends UncheckedBaseModel and is auto-generated by the Fern API definition toolchain.

Usage

Use Generation when working with the non-streaming Cohere Generate API response. This is the response type returned by co.generate() when stream is not set or is set to False.

Code Reference

Source Location

• Repository: Cohere Python SDK
• File: src/cohere/types/generation.py

Signature

class Generation(UncheckedBaseModel):
    id: str
    prompt: typing.Optional[str] = None
    generations: typing.List[SingleGeneration]
    meta: typing.Optional[ApiMeta] = None

Import

from cohere.types import Generation

I/O Contract

Fields

Field	Type	Required	Default	Description
id	str	Yes	--	Unique identifier for the generation request
prompt	Optional[str]	No	None	The prompt used for generation
generations	List[SingleGeneration]	Yes	--	List of generated results; each contains id, text, optional index, optional likelihood, and optional token_likelihoods
meta	Optional[ApiMeta]	No	None	API metadata including token counts and warnings

SingleGeneration Fields (Nested)

Field	Type	Required	Default	Description
id	str	Yes	--	Unique identifier for this individual generation
text	str	Yes	--	The generated text output
index	Optional[int]	No	None	The generation index (present when num_generations > 1)
likelihood	Optional[float]	No	None	Average log-likelihood of the generated text (when return_likelihoods is set)
token_likelihoods	Optional[List[SingleGenerationTokenLikelihoodsItem]]	No	None	Per-token log-likelihoods (when return_likelihoods is set)

Usage Examples

Basic Text Generation

import cohere

co = cohere.Client()

response = co.generate(
    prompt="Write a tagline for a coffee shop:",
    model="command",
    max_tokens=50,
)

print(f"Request ID: {response.id}")
print(f"Prompt: {response.prompt}")
print(f"Generated text: {response.generations[0].text}")

Multiple Generations

import cohere

co = cohere.Client()

response = co.generate(
    prompt="Suggest a name for a new programming language:",
    model="command",
    max_tokens=30,
    num_generations=5,
)

print(f"Generated {len(response.generations)} options:")
for gen in response.generations:
    print(f"  [{gen.index}] {gen.text.strip()}")

Generation with Likelihoods

import cohere

co = cohere.Client()

response = co.generate(
    prompt="The capital of France is",
    model="command",
    max_tokens=10,
    return_likelihoods="GENERATION",
)

gen = response.generations[0]
print(f"Text: {gen.text}")
print(f"Average likelihood: {gen.likelihood}")

if gen.token_likelihoods:
    for token_info in gen.token_likelihoods:
        print(f"  Token: {token_info.token!r} -> likelihood: {token_info.likelihood}")

Related Pages

• Environment:Cohere_ai_Cohere_python_Python_SDK_Runtime