OpenAI Delivers GPT-5.6 Sol Ultrafast Mode with Cerebras Hardware

TL;DR. OpenAI introduced "Ultrafast" mode for GPT-5.6 Sol, achieving 750 output tokens per second, powered by Cerebras hardware. - The new mode provides speed comparable to smaller models with the reasoning capabilities of a large language model. - This service is initially available to select OpenAI API customers, with broader access planned. - OpenAI is pricing inference speed as a tiered product, similar to cloud provider models.

• OpenAI releases "Ultrafast" mode for GPT-5.6 Sol, delivering up to 750 output tokens per second.
• Cerebras hardware, from a 0 billion partnership, provides the inference acceleration.
• Ultrafast mode aims to combine small model speed with large reasoning model capabilities.
• OpenAI implements a tiered pricing structure, turning inference speed into a distinct product.

Back to QLANKR News