A new preview mode targets enterprise users who need low-latency responses from OpenAI's most powerful model.
OpenAI is previewing 'Ultrafast,' a mode that runs GPT-5.6 Sol at 14 times its normal speed. The launch is aimed squarely at enterprise use cases where latency, not just model capability, determines whether AI can be embedded into real-time workflows.
The speed gains reportedly involve infrastructure partnerships, including work with Cerebras, to push inference throughput well beyond standard GPU serving. It's a preview, meaning broad availability and pricing details are still to come.
Latency has been the quiet blocker for enterprise AI in customer-facing and real-time applications. A 14x speed-up, if it holds in production, changes what use cases are commercially viable and puts pressure on rivals to match inference speed, not just benchmark scores.
The daily signal, curated. Get it in your inbox.
Subscribe on LinkedIn →