Supported Models

Find the right model for your workload. Compare hardware support across text, vision, reranking, generation, and recommendation, then open a deployment guide to get started.

Quick start → Hardware setup → Launch a service →

✓ Supported × Not supported — Not documented

LLM

Language models for chat, reasoning, and text generation.

Model / family	Ascend NPU	Cambricon MLU	Hygon DCU	Deployment guide
DeepSeek-V3 / R1 / V3.1	✅	✅	❌	Not yet available
DeepSeek-V3.2	✅	✅	❌	Not yet available
DeepSeek-V4	✅	✅	—	Guide ↗
Qwen2 / Qwen2.5 / QwQ	✅	✅	✅	Not yet available
Qwen3	✅	✅	✅	Not yet available
Qwen3 MoE	✅	✅	✅	Not yet available
Qwen3.5	✅	✅	—	Guide ↗
Kimi-K2	✅	❌	❌	Not yet available
Kimi-K2.5 / Kimi-K2.6	✅	—	—	Guide ↗
Llama2 / Llama3	✅	❌	❌	Not yet available
GLM-4.5	✅	❌	❌	Not yet available
GLM-4.6	✅	❌	❌	Not yet available
GLM-4.7	✅	❌	❌	Not yet available
GLM-5	✅	✅	❌	Guide ↗
GLM-5.1 / GLM-5.2 / GLM-5.3	✅	✅	—	Guide ↗
GLM-5.3-Flash	✅	—	—	Preview guide ↗
MiniMax-M2.7	✅	—	—	Preview guide ↗

VLM

Vision-language models for understanding images and text.

Model / family	Ascend NPU	Cambricon MLU	Hygon DCU	Deployment guide
MiniCPM-V	✅	❌	❌	Not yet available
MiMo-VL	✅	❌	❌	Not yet available
Qwen2.5-VL	✅	✅	✅	Not yet available
Qwen3-VL	✅	✅	✅	Not yet available
GLM-4.6V	✅	❌	❌	Not yet available
GLM-5.3-Flash-VL	✅	—	—	Preview guide ↗
VLM-R1	✅	❌	❌	Not yet available

Rerank

Models for scoring and reordering retrieved documents.

Model / family	Ascend NPU	Cambricon MLU	Hygon DCU	Deployment guide
Qwen3-Reranker	✅	❌	❌	Not yet available

DiT

Diffusion models for image and video generation.

Model / family	Ascend NPU	Cambricon MLU	Hygon DCU	Deployment guide
Flux	✅	❌	❌	Not yet available
Flux2	✅	—	—	Guide ↗
Wan2.2	✅	—	—	Guide ↗

Rec

Models for generative recommendation workloads.

Model / family	Ascend NPU	Cambricon MLU	Hygon DCU	Deployment guide
OneRec	✅	❌	❌	Not yet available
Qwen2	✅	❌	❌	Not yet available
Qwen2.5	✅	❌	❌	Not yet available
Qwen3	✅	❌	❌	Not yet available