MMOPD/Qwen3-4B-OT3-tau2-2ep

TEXT GENERATIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:4BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Sep 17, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

MMOPD/Qwen3-4B-OT3-tau2-2ep is a 4 billion parameter Qwen3-based causal language model developed by MMOPD, specifically fine-tuned as a tool-use (τ²-bench) domain teacher. This model excels in agentic customer service dialogues, particularly in airline, retail, and telecom scenarios, by generating assistant turns that include thinking processes and tool calls. It is an epoch-2 checkpoint from a supervised fine-tuning run on the inclusionAI/AReaL-tau2-data, optimized for robust tool-use capabilities within a 32,768-token context.

Loading preview...

Overview

MMOPD/Qwen3-4B-OT3-tau2-2ep is a 4 billion parameter Qwen3-based model, serving as the tool-use (τ²-bench) domain teacher within the MMOPD study. It is the epoch-2 checkpoint from a supervised fine-tuning (SFT) process, specifically trained on agent turns from the inclusionAI/AReaL-tau2-data dataset. This dataset comprises customer-service dialogues across airline, retail, and telecom domains, featuring tool calls and internal thinking processes.

Key Capabilities

  • Advanced Tool Use: Fine-tuned to generate assistant responses that include both internal thinking (<think>...</think>) and external tool calls (<tool_call>...<tool_call>), crucial for agentic applications.
  • Domain Specialization: Optimized for customer service interactions in telecom, airline, and retail sectors, handling complex dialogue contexts.
  • High Context Length: Supports a substantial context window of 32,768 tokens, enabling processing of longer conversations and detailed instructions.
  • Performance on τ²-bench: Achieves a pass^1 score of 57.9% on the τ²-bench telecom official prompt and 63.2% on the AReaL prompt, demonstrating strong performance in tool-use scenarios.

Good for

  • Developing AI Agents: Ideal for building and evaluating AI agents that require sophisticated tool-use capabilities in customer service or similar dialogue-driven environments.
  • Tool-Augmented LLM Applications: Suitable for applications where an LLM needs to reason, plan, and execute actions via external tools.
  • Research in Agentic AI: Provides a specialized model for studying and advancing tool-use and agentic behavior in language models.

Finetunes
1 models