According to CoinMeta, OpenAI has announced that the speed of its GPT-5.6 SOL model has been increased to 750 token per second, which is 14 times faster than the standard mode, and is exclusively available for corporate customers. This mode is powered by Cerebras and is currently only open to a limited number of API customers. OpenAI has already tested the ultrafast mode in scenarios such as troubleshooting, research, customer service, financial analysis, and development agent. The person in charge of Codex, Tibo Sottiaux, expects that this speed may become the default level within the next 1 to 2 years. Although the maximum output speed of the model is 14 times faster, in practical applications, the overall speed increase is usually between 3 and 10 times, depending on the complexity of the task.