OpenAI launches Ultrafast mode for GPT-5.6 Sol
Analysis based on 10 articles · First reported Aug 13, 2026 · Last updated Aug 14, 2026
The launch strengthens OpenAI's competitive position in the enterprise AI market by offering significantly faster inference, potentially attracting latency-sensitive customers and increasing API usage. Cerebras Systems gains a high-profile partnership that could boost its credibility and demand for its AI hardware, while competitors like Anthropic and Moonshot AI may face pressure to match performance.
On August 13, 2026, OpenAI announced a limited preview of Ultrafast, a new service tier for its GPT-5.6 Sol model, which runs up to 14 times faster than standard processing. Powered by Cerebras Systems hardware, Ultrafast generates up to 750 output tokens per second, targeting enterprise applications requiring real-time AI responses such as voice, customer support, financial research, security, and commerce. The feature is initially available through the OpenAI API to select customers, with broader access planned as capacity grows. OpenAI is testing Ultrafast with early customers including Jane Street Group, Podium, Basis, and Rogo, and is also using it internally for incident response and research. The launch intensifies competition with other AI labs like Anthropic and Moonshot AI, which offer faster or open-weight models.
Set up alerts, explore entity relationships, search across thousands of events, and build custom intelligence feeds.
Open Dashboard