agentscope-ai/QwenPaw-Flash-9B
QwenPaw-Flash-9B is a 9 billion parameter causal language model developed by AgentScope Team, fine-tuned from Qwen3.5. It is deeply optimized for autonomous agent scenarios, excelling in tool invocation, command execution, memory management, and multi-step planning within the QwenPaw ecosystem. This model features a native context length of 262,144 tokens and is designed for enhanced agentic performance with lower resource requirements.
Loading preview...
QwenPaw-Flash-9B: Optimized for Autonomous Agents
QwenPaw-Flash-9B is a 9 billion parameter causal language model, fine-tuned from Qwen3.5, specifically designed for the QwenPaw autonomous agent scenario. Developed by the AgentScope Team, this model is deeply optimized for agentic tasks, leveraging extensive, high-quality agent trajectory data from real QwenPaw environments.
Key Capabilities
- Active Memory Management: Autonomously identifies, stores, and retrieves user preferences and task states for logical consistency across multi-turn interactions.
- Native File Parsing: Excels at generating precise CLI commands and executing complex, multi-step file I/O tasks for terminal operations.
- Efficient Information Search: Enhanced for web-search tool invocation, featuring precise search intent recognition and multi-step web navigation.
- Intelligent Guidance: Built-in awareness of the QwenPaw feature map, proactively suggesting functional paths and troubleshooting based on real-time operational context.
Model Architecture and Performance
QwenPaw-Flash-9B shares architectural parameters with Qwen3.5-9B and includes a vision encoder. It boasts a native context length of 262,144 tokens. Benchmarks tailored to the QwenPaw environment demonstrate substantial improvements across multiple task categories, achieving performance comparable to leading flagship models while maintaining significantly lower resource requirements. This makes it ideal for complex agentic workflows requiring high logical consistency and efficient tool use.