Qwen3-235B-A22B
Open Weights
Alibaba Cloud
Alibaba's flagship open-weight Qwen3 model (April 2025), a 235B-parameter mixture-of-experts that activates 22B parameters per token. Trained on ~36T tokens across 119 languages under Apache 2.0, it supports seamless switching between a thinking mode for hard reasoning and a fast non-thinking mode, and is competitive with DeepSeek-R1, o1, o3-mini, and Gemini 2.5 Pro.
Strengths
- Flagship open-weight Qwen3 MoE under permissive Apache 2.0
- 235B parameters with only 22B active per token
- Switchable thinking / non-thinking modes
- Trained on ~36T tokens across 119 languages
- Competitive with DeepSeek-R1, o1, o3-mini, and Gemini 2.5 Pro
Caveats
- Native 32K context requires YaRN scaling to reach 131K
- Text-only (no vision or audio input)
- MoE serving needs substantial GPU memory
- Vendor
- Alibaba Cloud
- Context Window
- 131,072 tokens
- License
- Apache-2.0
- Release Date
- 2025-04-29
- Modalities
- text
Capabilities
Vision
Audio
Video
Tool Use
Resources
Pricing
Open weights under Apache 2.0 - free to download and self-host; hosted API providers vary. Native 32K context extends to 131K via YaRN.