Cerebras Code
Stop waiting on your model.·cerebrascoder.com
Cerebras Code is a platform that provides fast AI-powered code generation through the GLM 4.7 model, delivering completions at over 1,000 tokens per second. It can be used with any AI-friendly code editor or agent that accepts an API key, including Cline, OpenCode, and Crush. The product offers tiered token-based plans for different levels of coding usage, from casual demos to full-time development workflows.
What it's for
Features 11
- #1 ranking for tool calling on the Berkeley Function Calling Leaderboard
- Daily token allowance per plan tier
- GLM 4.7 model for code generation
- High-context completions
- Out-of-the-box integration with Cline, OpenCode, and Crush
- Over 1,000 tokens per second inference speed
- Performance comparable to Sonnet 4.5 in web-dev tasks
- Suited for agentic workflows
- Suited for IDE integrations and code refactoring
- Suited for multi-agent systems
- Works with any AI-friendly code editor or agent via API key
Pricing
All paid plans (Pro and Max) are currently sold out. Users are invited to follow on X and subscribe to a newsletter to be notified when new plans become available.
Currently marked sold out on the pricing page.
Limited tokens and requests
Currently marked sold out on the pricing page.
Up to 24 million tokens per day, enough for 3–4 hours of uninterrupted vibe coding
Currently marked sold out on the pricing page.
Up to 120 million tokens per day