Stay updated on Ultrafast mode

Sign up to receive updates on Ultrafast mode, including when capacity and access expands.

Ultrafast at a glance

• Powered by Cerebras, GPT‑5.6 Sol on Ultrafast generates up to 750 output tokens per second—up to 14× faster than Standard mode.

• Designed for live or near-production workloads like real-time voice, support, commerce, developer agents, financial research, and security response.

• Complete the form to be notified once access expands. Capacity is limited. We’ll evaluate customer inclusion based on workload fit and availability.