OpenAI Launches Ultra-Fast Mode for GPT-5.6 Sol, 14x Faster Output

OpenAI announced on August 14 the preview launch of an ultra-fast mode for its flagship AI model, GPT-5.6 Sol, offering output speeds 14 times faster than standard, generating up to 750 tokens per second. The mode is powered by Cerebras and aims to boost response times in user workflows.

OpenAI noted that speed and capability were previously a trade-off, but this mode combines both. Early use cases include incident response, financial research, customer service, e-commerce queries, and real-time experiments. The mode is currently in limited preview for select customers to assess its value in real-time operations.