linglingdan/DRIFT-8B-ToolUse

TEXT GENERATIONConcurrent Unit Cost:1Model Size:8BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Jun 17, 2026License:otherArchitecture:Transformer Featherless Exclusive Cold

DRIFT-8B-ToolUse is an 8 billion parameter causal language model fine-tuned by LianjiaTech based on Qwen/Qwen3-8B, designed for tool use applications. It features a context length of 32768 tokens and is optimized for generating tool calls. This model is particularly suited for tasks requiring interaction with external functions or APIs.

Loading preview...

DRIFT-8B-ToolUse: A Tool-Oriented Language Model

DRIFT-8B-ToolUse is an 8 billion parameter language model developed by LianjiaTech, fine-tuned from the Qwen/Qwen3-8B base model. It is specifically designed and optimized for tool use capabilities, enabling it to generate structured calls to external functions or APIs.

Key Capabilities

  • Tool Call Generation: Excels at understanding user intent and formulating appropriate tool calls.
  • Qwen3-8B Base: Leverages the robust architecture and performance of the Qwen3-8B model.
  • Extended Context Window: Supports a context length of 32768 tokens, allowing for processing longer prompts and complex tool use scenarios.
  • Bfloat16 Precision: Utilizes bfloat16 for efficient inference.

Limitations

  • Hallucination Risk: The model may occasionally hallucinate tool calls or produce invalid arguments.
  • Output Sensitivity: Performance is dependent on the serving template and the formatting of the tool schema.
  • Safety and Bias: Comprehensive documentation on safety, bias, and domain-specific failure modes is not provided.

Good For

  • Agentic Workflows: Ideal for building AI agents that need to interact with external systems.
  • Function Calling: Applications requiring the model to invoke specific functions based on user input.
  • Automated Task Execution: Scenarios where an LLM needs to orchestrate actions through tools.

Finetunes
1 models