[KongchangAI]
Modelstealth/cat-jam

Cat Jam

An inference model released in stealth mode on AI Gateway, aimed at coding and agentic tasks, emphasizing low-latency inference, supporting function calling and structured outputs, offering five selectable reasoning effort levels, and accepting text input only.

Timeline (last 90 days)

Oct 17

Running `/models` in OpenCode brings up the Select Model list, and in the video demonstration the list contained 9 free models: Big Pickle, Space Bunny, Cat Jam, Longcat 2.5 Preview, Step 5 Preview, Mimo v2.6 Flash, Muse Spark 1.3, and 2 Nvidia Nemotron models.

Unverified50%
Oct 15

Cat Jam is a reasoning model for programming and agentic tasks, focused on low-latency reasoning.

Unverified50%
Oct 15

Cat Jam supports function calling and structured output, and returns both reasoning text and the final answer.

Unverified50%
Oct 15

Cat Jam offers five levels of reasoning effort: low, medium, high, xhigh, and max.

Unverified50%
Oct 15

Cat Jam accepts only text input and does not support multimodal input such as images.

Unverified50%
Oct 15

Cat Jam is a stealth model, free for all users during the stealth period.

Unverified50%
Oct 15

Cat Jam's access channels include the AI SDK, OpenAI-compatible Chat Completions and Responses APIs, the Anthropic-compatible Messages API, and coding agents connected to AI Gateway.

Unverified50%
Oct 15

When calling Cat Jam, the model name stealth/cat-jam is used.

Unverified50%
Oct 15

Whether Cat Jam will become paid after the free period ends and how it will be priced are currently not stated officially.

Unverified50%
Oct 15

Cat Jam does not provide zero data retention (ZDR), and the official explanation states that prompts may be retained for security reasons.

Unverified50%

All Facts (10)

Source Articles