swarmauri_llm_cerebras 0.11.0.dev1

Provider-specific Swarmauri import package for Cerebras Inference chat completions with sync, async, streaming, batch, and JSON workflows.

Swarmauri Cerebras LLM

swarmauri_llm_cerebras provides the CerebrasModel import path for Swarmauri applications that use Cerebras Inference chat completions. The package re-exports the maintained implementation from swarmauri_standard, giving projects a provider-specific dependency for Cerebras-hosted Llama, Qwen, and GPT-OSS style model workflows.

Why Swarmauri Cerebras LLM?

Use this package when you want Cerebras Inference behind the Swarmauri LLMBase interface. CerebrasModel formats Swarmauri conversations for the Cerebras chat completions endpoint, supports sync and async prediction, streams response deltas, runs sequential and concurrent batches, captures usage data, and exposes model options such as temperature, top-p, seed, stop sequences, and JSON output mode.

FAQ

Q: Which Cerebras endpoint does this adapter call?

A: The adapter posts to https://api.cerebras.ai/v1/chat/completions and uses bearer-token authentication with a Cerebras API key.

Q: Which model names are configured locally?

A: The current configured list includes llama-4-scout-17b-16e-instruct, llama3.1-8b, llama-3.3-70b, llama-4-maverick-17b-128e-instruct, qwen-3-32b, qwen-3-coder-480b, and gpt-oss-120b. The default is llama3.1-8b.

Q: Does this package contain a separate Cerebras implementation?

A: No. It re-exports CerebrasModel from swarmauri_standard.llms.CerebrasModel so applications can depend on this provider package while sharing the common implementation.

Q: How do I request JSON output?

A: Pass enable_json=True to predict(), apredict(), stream(), or astream(). The adapter sets response_format to json_object.

Features

• Provider-specific import: from swarmauri_llm_cerebras import CerebrasModel.
• Swarmauri Conversation, HumanMessage, SystemMessage, and AgentMessage integration.
• Sync predict() and async apredict() chat completion calls.
• Sync stream() and async astream() streaming deltas.
• Sync batch() and async abatch() helpers.
• Configurable temperature, max_completion_tokens, top_p, seed, stop, and enable_json.
• Usage metadata mapping to Swarmauri UsageData when provider usage data is returned.
• Model discovery through the Cerebras /v1/models endpoint.
• Python 3.10, 3.11, 3.12, 3.13, and 3.14 support.

Prerequisites

• Cerebras Inference API key.
• Network access to api.cerebras.ai.
• Swarmauri conversations built from Swarmauri message classes.

Installation

Install with uv:

uv add swarmauri_llm_cerebras

Install with pip:

pip install swarmauri_llm_cerebras

Usage

Run a chat completion:

import os

from swarmauri_llm_cerebras import CerebrasModel
from swarmauri_standard.conversations.Conversation import Conversation
from swarmauri_standard.messages.HumanMessage import HumanMessage
from swarmauri_standard.messages.SystemMessage import SystemMessage

conversation = Conversation()
conversation.add_message(SystemMessage(content="Answer with concise technical prose."))
conversation.add_message(HumanMessage(content="What is Cerebras Inference?"))

model = CerebrasModel(
    api_key=os.environ["CEREBRAS_API_KEY"],
    name="llama3.1-8b",
)
result = model.predict(
    conversation=conversation,
    temperature=0.2,
    max_completion_tokens=256,
)

print(result.get_last().content)

Stream response deltas:

import os

from swarmauri_llm_cerebras import CerebrasModel
from swarmauri_standard.conversations.Conversation import Conversation
from swarmauri_standard.messages.HumanMessage import HumanMessage

conversation = Conversation()
conversation.add_message(HumanMessage(content="Write a short product summary."))

model = CerebrasModel(api_key=os.environ["CEREBRAS_API_KEY"])

for token in model.stream(conversation=conversation, max_completion_tokens=180):
    print(token, end="")

Use the async API:

import asyncio
import os

from swarmauri_llm_cerebras import CerebrasModel
from swarmauri_standard.conversations.Conversation import Conversation
from swarmauri_standard.messages.HumanMessage import HumanMessage


async def main() -> None:
    conversation = Conversation()
    conversation.add_message(HumanMessage(content="Return JSON with a status field."))

    model = CerebrasModel(api_key=os.environ["CEREBRAS_API_KEY"])
    result = await model.apredict(conversation=conversation, enable_json=True)

    print(result.get_last().content)


asyncio.run(main())

Related Packages

LLM provider packages:

• swarmauri_llm_ai21
• swarmauri_llm_anthropic
• swarmauri_llm_cohere
• swarmauri_llm_deepinfra
• swarmauri_llm_groq
• swarmauri_llm_mistral
• swarmauri_llm_openai

Foundational packages:

• swarmauri_core defines core interfaces.
• swarmauri_base provides base component classes.
• swarmauri_standard contains the shared CerebrasModel implementation.
• swarmauri provides namespace imports and plugin discovery.

Best Practices

• Store CEREBRAS_API_KEY in a secret manager or environment variable.
• Set name explicitly for reproducible model behavior.
• Use seed when deterministic replay matters for supported models.
• Use async methods for high-concurrency workloads.
• Validate JSON mode outputs before feeding them into strict downstream parsers.

License

Apache-2.0