Browsing: AI inference

OpenAI’s new Ultrafast mode runs GPT-5.6 Sol up to 14 times faster using Cerebras wafer-scale chips instead of Nvidia GPUs. Here’s what it means for the AI speed race.