Model•active
Qwen3-235B-A22B
Qwen3-235B-A22B is the large open-weight Qwen3 mixture-of-experts model with 235B total parameters and 22B active parameters.
Official siteVerified Oct 4, 2026
Overview
It is the flagship-scale representative in iilib's Qwen3 pilot and supports multilingual, reasoning, coding and agentic workloads with a 128K context.
Model specifications
- Family
- Qwen3
- Context
- 128,000 tokens
- Input
- Text
- Output
- Text
- Open weights
- Yes
- Capabilities
- Reasoning, Tool Use, Function Calling