zenlm/zen-agent-4b

Hugging Face
TEXT GENERATIONConcurrent Unit Cost:1Model Size:4BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Sep 28, 2025License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Warm

The zenlm/zen-agent-4b is a compact 4 billion parameter agent model developed by Hanzo AI and the Zoo Labs Foundation. Fine-tuned from Qwen3-4B-Instruct-2507, it features a dense Qwen3 architecture and a 256K-token context length. This model is specifically optimized for strong function calling and tool-use capabilities, making it suitable for agentic applications.

Loading preview...

Zen Agent 4b: Compact Agentic Model

Zen Agent 4b is a 4 billion parameter model developed by Hanzo AI and the Zoo Labs Foundation, specifically engineered for agentic applications. It is fine-tuned from the Qwen3-4B-Instruct-2507 base model, inheriting its dense Qwen3 architecture and an impressive 256K-token context window.

Key Capabilities

  • Strong Function Calling: Designed to excel at interpreting and executing function calls.
  • Enhanced Tool Use: Optimized for effective interaction with external tools and APIs.
  • Agentic Data Training: Benefits from specialized training data focused on agentic behaviors and interactions.
  • Large Context Window: Supports complex, multi-turn conversations and tasks with its 256K-token context.

Good For

  • Developing AI agents that require robust function calling.
  • Applications needing advanced tool-use capabilities.
  • Scenarios where a compact yet powerful agent model with a large context is beneficial.

Zen Agent 4b integrates Hanzo's identity training, agentic-data fine-tuning, and abliteration techniques, building upon the Apache 2.0 licensed Qwen3 base model.