OpenAI Launches Ultrafast Feature — GPT-5.6 Sol Is 14 Times Faster

OpenAI has released a new feature called Ultrafast, making its most powerful model, GPT-5.6 Sol, run 14 times faster than usual.

Key points to know:
• Speeds of up to 750 tokens per second — text streams out almost instantaneously
• Powered by chips from Cerebras, OpenAI’s new partner
• Currently in Preview — initially available to a select group of customers, with wider access coming later
• Target use cases: Customer service, financial market analysis, and e-commerce

[Insight: Previously, if you wanted speed, you had to switch to smaller, less capable models. Ultrafast eliminates that trade-off — delivering high performance and speed at the same time. Regular ChatGPT users in Cambodia will need to wait a bit longer before this speed becomes available.]