6 Sol UltraFast is previewed for enterprises, running 14 times faster, up to 750 tokens per second OpenAI Recently launched to enterprises GPT-5

GPT-5.6 Sol UltraFast is previewed for enterprises. The running speed is increased by 14 times, up to 750 tokens per second. OpenAI GPT-5.6 Sol UltraFast is previewed for enterprises. In this mode, the model running speed will be increased by up to 14 times, and a maximum of 750 tokens can be generated per second. However, the ultra-fast mode requires OpenAI approval based on enterprise scenarios. Enterprises can submit applications but whether they can be approved depends on OpenAI review. Suitable for workloads that require real-time response: OpenAI says that the ultra-fast mode is supported by Cerebras and is currently only applicable to the GPT-5.6 Sol model. After turning on the ultra-fast mode, the model can generate up to 750 tokens per second, which is equivalent to 14 times that of the standard mode. Of course, not all workloads require ultra-fast mode. OpenAI says that this mode is specially designed for workloads that require real-time or near-production environments, such as real-time voice, customer support, business, development of agents, financial research, and security research. Cerebras provides limited resources during the preview phase, so after an enterprise submits an application, OpenAI will decide whether to approve the use of ultra-fast mode based on the usage scenarios provided by customers. OpenAI will also gradually increase coverage based on workload. In the future, ultra-fast mode may support more enterprises through API calls. Interested corporate customers can click here to submit an application: via cnBeta.COM - Chinese Industry Information Station (author: Source: Blue Dot.com)