Pangu Embedded

聊天大模型

Pangu Embedded

Release date: 2025-06-30更新于: 2025-07-01 09:05:21961

Live demo

Parameters

70.0亿

Context length

32K

Chinese support

Supported

Reasoning ability

Pangu Embedded is an AI model published by 华为, released on 2025-06-30, for 聊天大模型, with 70.0B parameters, and 32K tokens context length, requiring about 14GB storage, under the Pangu Model License Agreement Version 1.0 license.

Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology

Pangu Embedded

Model basics

Reasoning traces

Not supported

Thinking modes

Thinking modes not supported

Context length

32K tokens

Max output length

32768 tokens

Model type

聊天大模型

Release date

2025-06-30

Model file size

14GB

MoE architecture

Total params / Active params

70.0B / N/A

Knowledge cutoff

No data

Pangu Embedded

Open source & experience

Code license

CC BY-SA-4.0

Weights license

Pangu Model License Agreement Version 1.0- 免费商用授权

GitHub repo

https://gitcode.com/ascend-tribe/pangu-embedded

Hugging Face

https://gitcode.com/ascend-tribe/pangu-pro-moe-model

Live demo

No live demo

Pangu Embedded

API details

API speed

3/5

No public API pricing yet.

Pangu Embedded

Benchmark Results

Pangu Embedded currently shows benchmark results led by GPQA (2 / 13, score 68), GSM8K (4 / 26, score 95.98), AIME 2024 (24 / 62, score 81.90). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.