加载中...

Hunyuan-A13B-Instruct

Name: Hunyuan-A13B-Instruct
Availability: InStock
Author: 腾讯AI实验室

Release date: 2025-06-27更新于: 2025-06-27 16:16:121,012

Live demo GitHub Hugging Face Compare

Parameters

800.0亿

Context length

256K

Chinese support

Supported

Reasoning ability

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

Hunyuan-A13B-Instruct

Model basics

Reasoning traces

Not supported

Thinking modes

Thinking modes not supported

Context length

256K tokens

Max output length

32000 tokens

Model type

基础大模型

Release date

2025-06-27

Model file size

160.8 GB

MoE architecture

Total params / Active params

800.0B / N/A

Knowledge cutoff

No data

Hunyuan-A13B-Instruct

Open source & experience

Code license

TENCENT HUNYUAN COMMUNITY LICENSE AGREEMENT

Weights license

TENCENT HUNYUAN COMMUNITY LICENSE AGREEMENT- 免费商用授权

GitHub repo

https://github.com/Tencent-Hunyuan/Hunyuan-A13B

Hugging Face

https://huggingface.co/tencent/Hunyuan-A13B-Instruct

Live demo

https://hunyuan.tencent.com/?model=hunyuan-a13b

Hunyuan-A13B-Instruct

Official resources

Paper

Hunyuan-A13B

DataLearnerAI blog

腾讯开源Hunyuan-A13B大模型：MoE架构，混合推理（支持直接回复和带推理过程后回复），原WizardLM团队打造，评测结果超Qwen2.5-72B和Qwen3-A22B

Hunyuan-A13B-Instruct

API details

API speed

3/5

No public API pricing yet.

Hunyuan-A13B-Instruct

Benchmark Results

Hunyuan-A13B-Instruct currently shows benchmark results led by MATH (3 / 41, score 94.30), AIME 2024 (17 / 62, score 87.30), MMLU (19 / 62, score 88.17). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.