Supported Model List

View Source On Gitee

Model

Status

Backend Supported

Hardware Supported

Model Download Link

DeepSeek-V3

Supported

MindFormers

Atlas 800I A2

DeepSeek-V3

DeepSeek-R1

Supported

MindFormers

Atlas 800I A2

DeepSeek-R1

DeepSeek-R1 W8A8

Supported

MindFormers

Atlas 800I A2

DeepSeek-R1-W8A8

DeepSeek-R1 W8A4

Supported

MindFormers

Atlas 800I A2

DeepSeek-R1-W8A4

Telechat2

Supported

MindFormers

Atlas 800I A2

TeleChat2-7B-32K, TeleChat2-35B-32K

GLM-4.5

Supported

MindFormers

Atlas 800I A2

GLM-4.5, GLM-4.5-Air

GLM-4

Supported

MindFormers

Atlas 800I A2

GLM-4-9B-0414GLM-4-32B-0414

Qwen2.5

Supported

Native, MindFormers

Atlas 800I A2, Atlas 300I Duo(Testing)

Qwen2.5-0.5B-Instruct, Qwen2.5-1.5B-Instruct, Qwen2.5-3B-Instruct, Qwen2.5-7B-Instruct, Qwen2.5-14B-Instruct, Qwen2.5-32B-Instruct, Qwen2.5-72B-Instruct

Qwen3

Supported

Native, MindFormers

Atlas 800I A2, Atlas 300I Duo

Qwen3-0.6B, Qwen3-1.7B, Qwen3-4B, Qwen3-8B, Qwen3-14B, Qwen3-32B

Qwen3-235B-A22B

Supported

Native, MindFormers

Atlas 800I A2

Qwen3-235B-A22B

Qwen3-30B-A3B

Testing

Native, MindFormers

Atlas 800I A2

Qwen3-30B-A3B

Qwen2.5-VL

Supported

Native

Atlas 800I A2

Qwen2.5-VL-3B-Instruct, Qwen2.5-VL-7B-Instruct, Qwen2.5-VL-32B-Instruct, Qwen2.5-VL-72B-Instruct

QwQ-32B

Testing

Native, MindFormers

Atlas 800I A2

QwQ-32B

Llama3.1

Testing

Native

Atlas 800I A2

Llama-3.1-8B-Instruct, Llama-3.1-70B-Instruct, Llama-3.1-405B-Instruct

Llama3.2

Testing

Native

Atlas 800I A2

Llama-3.2-1B-Instruct, Llama-3.2-3B-Instruct

Model Description

  1. User can refer to Environment Variable List, and set the model backend by environment variable VLLM_MS_MODEL_BACKEND.

  2. The native model backend currently supports the Qwen2.5, Qwen2.5VL, Qwen3 and Llama series models; the MindSpore Transformers backend supports Qwen, DeepSeek, TeleChat and GLM series models.

  3. 300I Duo has supported Qwen3 model, and other models are in the process of adaptation.