Cannae-AI/Gpt-oss-120B-Qwen3-Distill

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:4BQuant:BF16Ctx Length:32kPublished:Dec 9, 2025License:apache-2.0Architecture:Transformer0.0K Open Weights Warm

Cannae-AI/Gpt-oss-120B-Qwen3-Distill is a reasoning-distilled version of the Gpt-oss-120B model, developed by Cannae-AI. This model specializes in mathematical reasoning, having been fine-tuned using Qwen3-4B-Thinking-2507 and thousands of generated complete math reasoning processes and answers. It is optimized for tasks requiring logical deduction and step-by-step mathematical problem-solving.

Loading preview...

Cannae-AI/Gpt-oss-120B-Qwen3-Distill Overview

This model, developed by Cannae-AI, is a specialized large language model focused on enhancing reasoning capabilities, particularly in mathematics. It is a distilled version of the original gpt-oss-120b model.

Key Capabilities

  • Reasoning Distillation: The model has undergone a unique distillation process, leveraging Qwen3-4B-Thinking-2507.
  • Mathematical Problem Solving: It is specifically trained on thousands of generated complete math reasoning processes and answers, making it proficient in handling mathematical tasks.
  • Logical Deduction: Optimized for tasks that require step-by-step logical deduction and problem-solving.

Recommended Use Cases

  • Mathematical Reasoning: Ideal for applications requiring accurate and detailed mathematical problem-solving.
  • Educational Tools: Can be integrated into tools for teaching or assisting with math-related queries.
  • Automated Proof Generation: Potentially useful for generating or verifying mathematical proofs and logical sequences.

Inference Settings

For optimal performance, the recommended inference settings are:

  • temperature = 0.7
  • top_p = 0.8
  • top_k = 20

Finetunes
1 models