OpenAI Launches Ultrafast Processing Mode for GPT-5.6 Sol Model

OpenAI has announced the launch of the Ultrafast processing mode for the GPT-5.6 Sol model, enabling production of up to 750 tokens per second. This new mode achieves speeds up to 14 times faster than the standard processing mode, enhancing the model's efficiency without the need to switch to a smaller or less capable model. The new mode relies on the Wafer-Scale Engine architecture from Cerebras, which utilizes advanced memory to meet high processing demands. This technology represents a significant advancement in the models' ability to handle complex tasks more swiftly, opening new horizons for artificial intelligence applications.