katanemo/Arch-Agent-1.5B

Arch-Agent-1.5B: Advanced Function Calling LLM

katanemo/Arch-Agent-1.5B is a 1.5 billion parameter large language model from the Arch-Agent series, developed by katanemo. It is specifically engineered for sophisticated function calling and agent-based applications, enabling complex multi-step and multi-turn workflows. The model's core strength lies in its ability to manage intelligent tool selection, dynamic planning, and seamless integration with external APIs and services, ensuring reliability and precision across extended function call sequences.

Key Capabilities

• Multi-Turn Function Calling: Maintains contextual continuity for natural, ongoing conversations with nested or evolving tool use.
• Multi-Step Function Calling: Plans and executes sequences of function calls, adapting dynamically based on intermediate results and decomposing goals into sub-tasks.
• Agentic Capabilities: Provides advanced decision-making and workflow management for complex agentic tasks, including seamless tool coordination and error recovery.

Performance

The Arch-Agent series is evaluated on the Berkeley Function-Calling Leaderboard (BFCL), demonstrating strong performance against commonly-used models. For evaluation, Arch-Agent models utilize YaRN scaling and are assessed with a context length of 64K, highlighting their proficiency in handling extensive conversational contexts.

Good For

• Developing AI agents that require complex, multi-step task execution.
• Applications needing robust multi-turn dialogue management with integrated tool use.
• Scenarios demanding precise and reliable function calling for external API interactions.