Qwen3-235B-A22B is a 235B parameter mixture-of-experts (MoE) model developed by Qwen, activating 22B parameters per forward pass. It supports seamless switching between a "thinking" mode for complex reasoning, math, and...
Quick stats across all benchmark runs.
Models with similar or better quality but different tradeoffs
How this model performs across different benchmarks
Compare cost efficiency across all models
Performance trends across all benchmark runs
Number of benchmark runs over time
Get started with this model using OpenRouter
Compare performance with other models from the same creator