PAI-Model Gallery supports one-click deployment of Qwen3 full-size models on the cloud
Qwen3 large language model cloud deployment guide, one-click experience of AI efficiency and convenience.
Core content:
1. Qwen3 model features and breakthroughs in multiple fields
2. PAI-Model Gallery integrates Qwen3 model and provides zero-code deployment solution
3. Detailed steps and configuration guide for one-click deployment of Qwen3 model
01
Qwen3 is the latest generation of large language models in the Qwen series, providing a series of dense and mixture of experts (MOE) models. Based on extensive training, Qwen3 has made breakthrough progress in reasoning, command following, agent capabilities and multi-language support, with the following key features:
Unique support for seamless switching between thinking mode (for complex logical reasoning, mathematics, and coding) and non-thinking mode (for efficient general conversations), ensuring optimal performance in a variety of scenarios. Significantly enhanced reasoning capabilities , surpassing the previous QwQ (in thinking mode) and Qwen2.5 instruction model (in non-thinking mode) in mathematics, code generation, and common sense logic reasoning. Superior alignment with human preferences , excels in creative writing, role-playing, multi-turn conversations, and command following, providing a more natural, engaging, and immersive conversational experience. Expert in agent capabilities , can accurately integrate external tools in thinking and non-thinking modes, and leads in open source models in complex agent-based tasks. It supports more than 100 languages and dialects , and has powerful multi-language understanding, reasoning, command following and generation capabilities.
02
Model Gallery is a product component of Alibaba Cloud's artificial intelligence platform PAI. It integrates high-quality pre-trained models from domestic and foreign AI open source communities, covering various fields such as LLM, AIGC, CV, and NLP. Through PAI's adaptation of these models, users can implement the entire process from training to deployment to reasoning in a zero-code manner, simplifying the model development process and bringing faster, more efficient, and more convenient AI development and application experience to developers and enterprise users.
PAI-Model Gallery access address: https://pai.console.aliyun.com/#/quick-start/models
Alibaba Cloud PAI-Model Gallery has simultaneously connected all the open source models of Qwen3, providing enterprise-level deployment solutions.
✅Zero code one-click deployment
✅Automatically adapt to cloud resources
✅Out -of-the-box API
✅Full process operation and maintenance hosting
✅ Enterprise-level security data does not leave the domain
03
⬇️Experience it now⬇️
The following takes the Qwen3-8B model deployment as an example (the inference cost is low and can be used for fast verification).
The following table provides the minimum configuration required for deployment, as well as the maximum number of tokens supported on different inference frameworks when deployed on different machine models.
04
In addition to the full range of Qwen3 models, PAI-Model Gallery continues to provide rapid deployment, training, and evaluation practices for popular models in the open source community.
- DeepSeek-R1 inference performance optimized version
The inference performance is improved. Under the same latency constraint, the throughput can be increased by 492% . Under the same throughput condition, the first token latency is reduced by 86%, and the latency between tokens is reduced by 69%.
