Meta's Muse Spark Open Model Arrives on Vercel's AI Gateway
Vercel has added Meta's Muse Spark Open to its AI Gateway. The open-weight reasoning model is aimed at coding and agentic workflows and accepts both text and image input.
- Meta's open-weight reasoning model Muse Spark Open is now available on Vercel's AI Gateway.
- The model targets coding and agentic work such as multi-file refactoring, debugging, and understanding large codebases.
- It supports text and image input with a 1M-token context window.
- Developers call it with the model name meta/muse-spark-open via the AI SDK, OpenAI-compatible APIs, the Anthropic-compatible Messages API, or coding agents.
- According to Vercel, AI Gateway reflects provider pricing with no markup and no platform fee on inference.
What Muse Spark Open offers
According to Vercel, Muse Spark Open is an open-weight reasoning model from Meta designed for coding and agentic workflows. The listed use cases include multi-file refactoring, debugging, and understanding large codebases.
The model supports text and image input and has a 1M-token context window. Vercel says this allows agents to work with code and screenshots in the same request. Vercel also states that the model is available with ZDR (Zero Data Retention) and no training, and that developers do not need to create separate provider accounts.
How to access it
Developers use meta/muse-spark-open as the model name across all API formats and coding agents. Vercel lists the following supported access routes:
- The AI SDK
- The OpenAI-compatible Chat Completions and Responses APIs
- The Anthropic-compatible Messages API
- Coding agents connected to AI Gateway
In the AI SDK, the model is called through streamText with the model set to 'meta/muse-spark-open'. Vercel's example prompt asks the model to outline three steps for safely refactoring an authentication module.
To use the model in coding agents such as Claude Code, Codex, and Cursor, Vercel instructs users to install the latest Vercel CLI with pnpm i -g vercel@latest and run vercel ai-gateway coding-agents setup. After that, users select meta/muse-spark-open in the agent. Vercel also points to a coding agents guide for details and a model playground for trying the model.
About AI Gateway and what it means for developers
Vercel describes AI Gateway as a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. Features named in the announcement include built-in custom reporting, Zero Data Retention support, budgets for API keys, and routing rules.
On pricing, Vercel says AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. The announcement does not give per-token prices for Muse Spark Open itself.
For developers, the addition means an open-weight Meta model can be called through the same endpoint and API formats they may already use for other models, with one model name across tools. Teams that rely on coding agents can select it after running the CLI setup, without managing a separate provider account.
Vercel: Muse Spark Open now available on AI Gateway
