tepirale/Ornith-Agents-A1-3.6-35B-A3B-task_arithmetic

TEXT GENERATIONConcurrent Unit Cost:3Model Size:35.1BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Jul 3, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

The tepirale/Ornith-Agents-A1-3.6-35B-A3B-task_arithmetic model is a 35.1 billion parameter language model merged using the Task Arithmetic method, based on Qwen/Qwen3.5-35B-A3B. It integrates capabilities from deepreinforce-ai/Ornith-1.0-35B and InternScience/Agents-A1, featuring a 32768 token context length. This model is optimized for reasoning and tool-calling, supporting Qwen3-style reasoning content and XML-based tool calls. It excels in complex problem-solving and interactive agentic workflows.

Loading preview...

Model Overview

tepirale/Ornith-Agents-A1-3.6-35B-A3B-task_arithmetic is a 35.1 billion parameter language model created by merging existing pre-trained models using the Task Arithmetic method. Its base model is Qwen/Qwen3.5-35B-A3B, and it incorporates functionalities from deepreinforce-ai/Ornith-1.0-35B and InternScience/Agents-A1. This merge aims to combine their strengths, particularly in reasoning and agentic capabilities.

Key Capabilities

  • Advanced Reasoning: Leverages a Qwen3-style reasoning parser, enabling the model to generate explicit thought processes (<think>...</think>) before producing a final answer, enhancing transparency and problem-solving. This is activated via enable_thinking in chat template kwargs.
  • Tool-Calling: Supports Qwen3 XML-based tool-calling, allowing it to interact with external tools and APIs for more complex tasks.
  • Extended Context: Features a substantial context window of 32768 tokens, facilitating the processing of longer inputs and maintaining conversational coherence over extended interactions.
  • Optimized for Agentic Workflows: The integration of Ornith-1.0-35B and Agents-A1 suggests a strong focus on agent-like behaviors, decision-making, and multi-step task execution.

Good For

  • Complex Problem Solving: Ideal for applications requiring detailed reasoning and multi-step logical deduction.
  • Automated Agents: Suitable for building AI agents that need to plan, use tools, and interact with systems.
  • Interactive Applications: Benefits use cases where explicit reasoning traces are valuable for debugging or user understanding.
  • Research and Development: Provides a robust foundation for exploring advanced agentic AI and reasoning capabilities.

Finetunes
2 models