Qwen: Qwen3 235B A22B Instruct 2507
qwen
Qwen3-235B-A22B-Instruct-2507 is a multilingual, instruction-tuned mixture-of-experts language model based on the Qwen3-235B architecture, with 22B active parameters per forward pass. It is optimized for general-purpose text generation, including instruction following,...
- Context window
- 262,144 tokens
- Input cost
- $0.07 / 1M
- Output cost
- $0.10 / 1M
- Latency (p50)
- —