Technology
Qwen 3
Qwen 3: Alibaba Cloud's flagship, open-source (Apache 2.0) LLM series, featuring a hybrid Mixture-of-Experts (MoE) architecture and dual-mode reasoning for balancing speed and complexity.
This is Qwen 3 (Tongyi Qianwen), Alibaba Cloud's latest open-source language model series. It utilizes a hybrid architecture, including MoE variants like the Qwen3-235B-A22B (22B active parameters), to deliver high performance with efficiency. The core innovation is the Hybrid Reasoning Engine: developers can toggle between 'Thinking Mode' for complex, multi-step tasks (like coding or math) and 'Non-Thinking Mode' for rapid, direct responses. Qwen 3 is built for global deployment, supporting 119 languages and featuring a substantial 128K token context window across many of its models.
What builders pair with Qwen 3
Projects using both technologies. Select a pairing to see a project.
Pairing: Docker
[UofT] Print Anything: Building the Pipeline from Natural Language to 3D Object
Pairing: HuggingFace
[UofT] Print Anything: Building the Pipeline from Natural Language to 3D Object
Pairing: LoRA
[UofT] Print Anything: Building the Pipeline from Natural Language to 3D Object
Pairing: Three
[UofT] Print Anything: Building the Pipeline from Natural Language to 3D Object
Recent Talks & Demos
Showing 1-1 of 1
