LLM
Muse Spark Open
Muse Spark OpenMeta's 816B open-weight MoE model for coding agents, with hybrid attention for high throughput and long context.
Muse Spark Open is a model designed specifically for coding agents. It combines strong code generation and reasoning with reliable tool use, helping agents understand repositories, plan changes, write code, and complete complex software-engineering tasks.
Built to be served efficiently in production, Muse Spark Open supports responsive, high-throughput agentic workflows without sacrificing capability. It gives teams a practical foundation for building fast, reliable coding experiences at scale.
Input
1import os
2from openai import OpenAI
3
4client = OpenAI(
5 base_url="https://inference.baseten.co/v1",
6 api_key=os.environ["BASETEN_API_KEY"],
7)
8
9response = client.chat.completions.create(
10 model="meta-models/muse-spark-open",
11 messages=[
12 {
13 "role": "developer",
14 "content": "You are a senior backend engineer reviewing pull requests. For each code snippet the user shares, respond with: 1) a one-line summary of what the code does, 2) any bugs or issues, 3) one concrete improvement suggestion. Be direct. Do not repeat the code back.",
15 },
16 {
17 "role": "user",
18 "content": "def div(a, b):\n return a / b",
19 },
20 ],
21)
22
23print(response.model_dump_json(indent=2))
