Cat Jam lands on AI Gateway as a stealth reasoning model, free for coding and agent tasks
Synopsis
AI Gateway now offers Cat Jam, a stealth reasoning model free to use during the stealth period. It targets coding and agentic tasks with low-latency inference, supports function calling so agents can reach external tools, and returns structured outputs that follow a caller-supplied JSON schema. Responses carry the model's reasoning text alongside the answer. Input is text only, reasoning effort ranges from low to max, and Zero Data Retention is not available.
Interpretation
Cat Jam appears on AI Gateway as a stealth reasoning model that costs nothing during the stealth period and is aimed at coding and agentic tasks with low-latency inference. Other models on the gateway are billed at provider pricing, with no markup and no platform fee on inference; Cat Jam is offered free while in stealth. The free, stealth-period status comes directly from the release description, which does not say what happens after the stealth period ends.
Agents can connect Cat Jam to their own tools through function calling, and callers can require structured outputs that follow a JSON schema they supply, with reasoning text returned alongside the answer. Function calling lets the model invoke external tools, while structured outputs constrain the response to the given schema rather than free text. The description lists function calling and structured outputs, states text-only input, and names five reasoning efforts: low, medium, high, xhigh and max.
Access runs through the AI SDK, the OpenAI-compatible Chat Completions and Responses APIs, the Anthropic-compatible Messages API, and coding agents connected to AI Gateway. Using it inside Claude Code, Codex or Cursor requires installing the latest Vercel CLI, running setup, then selecting stealth/cat-jam in the agent. The model name is stealth/cat-jam, and the release supplies the command-line setup step and a coding agents guide as the access path.
Perspective
The release addresses developers using AI Gateway and describes model capabilities and access: text input, function calling, structured outputs following a supplied JSON schema, reasoning text in responses, and five reasoning efforts. The stated access paths are the AI SDK, the OpenAI-compatible Chat Completions and Responses APIs, the Anthropic-compatible Messages API, and coding agents such as Claude Code, Codex and Cursor calling stealth/cat-jam. Free use applies only during the stealth period.
Zero Data Retention is not available for this model, and prompts may be retained for safety reasons, a direct constraint for teams handling sensitive code or data. Input is text only. The release does not say when the stealth period ends, nor whether or how the model will be offered or priced afterward.