RecursiveMAS/Sequential-Light-Planner-Qwen3-1.7B

TEXT GENERATIONConcurrency Cost:1Model Size:2BQuant:BF16Ctx Length:32kPublished:Apr 28, 2026License:mitArchitecture:Transformer Open Weights Cold

RecursiveMAS/Sequential-Light-Planner-Qwen3-1.7B is a 1.7 billion parameter model developed by RecursiveMAS, based on the Qwen3 architecture with a 32K context length. It functions as a role-specific Planner Agent within the RecursiveMAS multi-agent framework, designed to produce initial reasoning plans for sequential refinement by other agents. This model is optimized for collaborative problem-solving in a multi-agent system rather than standalone general text generation.

Loading preview...

RecursiveMAS/Sequential-Light-Planner-Qwen3-1.7B Overview

This model, developed by RecursiveMAS, is a specialized Planner Agent within the novel RecursiveMAS multi-agent framework. Unlike general-purpose language models, it is not intended for standalone text generation but rather for a specific role in a collaborative system. Based on the Qwen3-1.7B architecture, this agent is designed to initiate reasoning processes within a "Sequential-Light" collaboration style.

Key Capabilities

  • Initial Plan Generation: Responsible for creating the foundational reasoning plan in a multi-agent workflow.
  • Latent-Space Recursion: Operates within a framework where agents iteratively exchange and refine latent states.
  • Role-Specific Functionality: Optimized exclusively for its planner role, passing its output to subsequent agents via lightweight RecursiveLink modules for further refinement.

When to Use This Model

  • Multi-Agent System Development: Ideal for researchers and developers building systems using the RecursiveMAS framework, particularly for the initial planning phase.
  • Collaborative AI Research: Suitable for exploring scalable agent collaboration through latent-space recursion.

For detailed usage and integration into the RecursiveMAS framework, refer to the official GitHub repository.