xxwu/Agent-STAR-RL-1.5B

Hugging Face
TEXT GENERATIONPricing:Input $0.04 / Cached $0.008 / Output $0.08Concurrent Unit Cost:1Model Size:1.5BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Mar 23, 2026License:mitArchitecture:Transformer Open Weights Featherless Exclusive Loading

The xxwu/Agent-STAR-RL-1.5B is a 1.5 billion parameter language model, built on the Qwen2.5-1.5B-Instruct backbone, developed by xxwu. It is specifically fine-tuned using Reinforcement Learning (RL) within the STAR pipeline (Data Synthesis → SFT → RL) for long-horizon tool orchestration and planning tasks. This model excels at handling complex, multi-turn environments, particularly benefiting from scale-aware RL recipes for smaller models. Its primary application is in enabling agents to effectively use tools and plan over extended interactions.

Loading preview...

Agent-STAR-RL-1.5B: Long-Horizon Tool-Using Agent

Agent-STAR-RL-1.5B is a 1.5 billion parameter model, based on the Qwen2.5-1.5B-Instruct backbone, specifically trained for long-horizon tool orchestration and planning. Developed by xxwu, this model is the result of the STAR pipeline (Data Synthesis → SFT → RL), a systematic study into reinforcement learning (RL) for tool-using agents.

Key Capabilities

  • Reinforcement Learning (RL) Optimization: Trained with RL to navigate complex, multi-turn environments.
  • Tool Orchestration: Designed to effectively use and coordinate various tools for task completion.
  • Long-Horizon Planning: Capable of planning and executing actions over extended interaction sequences.
  • Scale-Aware Training: Benefits from specialized RL recipes, including staged rewards and enhanced exploration, tailored for smaller models to manage complex constraints.

Good For

  • Developing Tool-Using Agents: Ideal for researchers and developers building agents that need to interact with external tools.
  • Complex Multi-Turn Environments: Suited for applications requiring agents to maintain context and plan across many steps.
  • Research in RL for LLMs: Provides a practical example and testbed for studying RL design spaces in language models, as detailed in the associated paper: Demystifying Reinforcement Learning for Long-Horizon Tool-Using Agents: A Comprehensive Recipe.

This model leverages a 32768 token context length and is demonstrated using the TravelPlanner testbed, with training data from the Agent-STAR-TravelDataset.

Finetunes
1 models