Cannae-AI
/

Gpt-oss-120B-Qwen3-Distill

Text Generation

Model card Files Files and versions

Gpt-oss-120B-Qwen3-Distill

Model Description

This is a reasoning distilled version of gpt-oss-120b using Qwen3-4B-Thinking-2507 and thousands of generated complete math reasoning processes and answers using gpt-oss-120b-4bit!

Inference settings : temperature = 0.7, top_p = 0.8, top_k = 20
License: apache-2.0
Developed by : Cannae-AI

GPT-OSS-120B

Downloads last month: 191

Safetensors

Model size

4B params

Tensor type

BF16

·

Model tree for Cannae-AI/Gpt-oss-120B-Qwen3-Distill

Base model

Qwen/Qwen3-4B-Thinking-2507

Finetuned

unsloth/Qwen3-4B-Thinking-2507

Finetuned

(78)

this model

Quantizations

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

Distilled LLms

This is a collection of various distilled llms by Cannae ai! • 3 items • Updated Dec 17, 2025