Cerebras OpenAI Basic

Compare synchronous, asynchronous, and streaming calls through Cerebras' OpenAI-compatible endpoint.

basic.py
"""
Cerebras Openai Basic
=====================

Cookbook example for `cerebras_openai/basic.py`.
"""

import asyncio

from agno.agent import Agent
from agno.models.cerebras import CerebrasOpenAI

# ---------------------------------------------------------------------------
# Create Agent
# ---------------------------------------------------------------------------

agent = Agent(
    model=CerebrasOpenAI(id="gpt-oss-120b"),
    markdown=True,
)

# Print the response in the terminal

# ---------------------------------------------------------------------------
# Run Agent
# ---------------------------------------------------------------------------
if __name__ == "__main__":
    # --- Sync ---
    agent.print_response("write a two sentence horror story")

    # --- Sync + Streaming ---
    agent.print_response("write a two sentence horror story", stream=True)

    # --- Async ---
    asyncio.run(agent.aprint_response("write a two sentence horror story"))

    # --- Async + Streaming ---
    asyncio.run(agent.aprint_response("write a two sentence horror story", stream=True))

Run the Example

Set up your virtual environment

uv venv --python 3.12
source .venv/bin/activate

Install dependencies

uv pip install -U agno cerebras-cloud-sdk openai

Export your Cerebras OpenAI-compatible API key

export CEREBRAS_API_KEY="your_cerebras_api_key_here"

Run the example

Save the code above as basic.py, then run:

python basic.py

Full source: cookbook/90_models/cerebras_openai/basic.py