Israel399B parameters256K contextEnterprise

Explore Jamba Large 1.7

Use AI21 Labs' large hybrid SSM-Transformer model for enterprise chat, long-document analysis, grounded answers, and tool-assisted workflows.

Model

ai21labs/AI21-Jamba-Large-1.7

Input

Configure your request

System prompt Optional

Set the model role and response constraints.

Grounding documents Optional

Optional documents as a JSON array for grounded generation.

Conversation history Optional

Optional previous messages as JSON.

Your message

Parameters

Stream responseMust be disabled when requesting multiple responses or using tools.

Temperature

AI21 API default.

Top P

Maximum output tokens

AI21’s Jamba API permits up to 4,096 output tokens.

Number of responses

Must be one when streaming.

Stop sequences

Response format

Function tools (JSON)

Tool choice