Cerebras models

Agent Serve currently tracks 6 Cerebras models including Qwen 3.8 27B, GPT OSS 120B, Llama 3 1 8B, Qwen 3 235B A22B Instruct 2507, Zai GLM 4.7, Gemma 4 31B.

GPT OSS 120B is the default Cerebras model in Agent Serve.

Llama 3 1 8B currently has the lowest listed input price at $0.1/1M tokens.

Qwen 3.8 27B currently has the largest listed context window at 131k tokens.

All Cerebras models in Agent Serve support tool use and function calling, allowing agents to invoke external APIs, query databases, and run custom actions.