Qwen 3.8 27B available on Cerebras at 1500 tok/SEC
Cerebras has released the Qwen 3.8 27B language model, offering a 27‑billion‑parameter variant that achieves about 1,500 tokens per second on its hardware. The model builds on the Qwen series, delivering faster inference for natural‑language tasks while maintaining competitive accuracy. Developers can access the new version through Cerebras’ platform, enabling high‑throughput applications in research and industry.