CharlieLLL/Qwen3-1.7B-BrowseComp-Worker-OPD-iter149-0917

TEXT GENERATIONPricing:Input $0.32 / Cached $0.064 / Output $1.6Concurrent Unit Cost:1Model Size:2BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Sep 19, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

CharlieLLL/Qwen3-1.7B-BrowseComp-Worker-OPD-iter149-0917 is a 1.7 billion parameter Qwen3-based model, specifically an evaluation export of a browsing worker. It is fine-tuned for complex web browsing and research tasks, demonstrated by its performance in BrowseComp and DR9K benchmarks when orchestrated with various larger LLMs. This model is designed to function as a specialized worker within a larger AI system, excelling at information retrieval and task execution in a browsing environment.

Loading preview...

Overview

CharlieLLL/Qwen3-1.7B-BrowseComp-Worker-OPD-iter149-0917 is a 1.7 billion parameter model based on the Qwen3 architecture, specifically an evaluation checkpoint from iteration 149. This model functions as a specialized browsing worker, fine-tuned using an OPD (Open-ended Problem-solving Dialogue) training run. It is designed to operate within an orchestrator-worker setup, rather than as a standalone chat model.

Key Capabilities and Performance

This model's primary strength lies in its ability to perform complex tasks requiring web browsing and information extraction. Its performance is evaluated in an orchestrated environment, showcasing its effectiveness when paired with larger LLMs acting as orchestrators. Key evaluation benchmarks include:

  • BrowseComp: Achieved scores up to 69.33% (104/150) when orchestrated by MiMo-V2.5.
  • DR9K: Achieved scores up to 59.38% (152/256) when orchestrated by Inkling-Small.

These results highlight its utility in multi-agent systems for tasks like web research and data collection. The model utilizes a Qwen/Qwen3-1.7B tokenizer and supports a context length of 40,960 tokens.

Intended Use

This model is specifically intended for use as a component in multi-agent AI systems where it can act as a browsing and information retrieval worker. It is not designed for direct, generic chat generation. Developers should integrate it with an orchestrator model and appropriate search tools to reproduce its evaluated performance.

Finetunes
1 models