OpenAI previews Ultrafast, a Cerebras-powered API tier that runs GPT-5.6 Sol up to 14x faster

Archive entry: this covers an event from August 13, 2026, and we published it on November 8, 2026. The details come from sources dated at the time and may have changed since.

On August 13, 2026, OpenAI previewed Ultrafast, a Cerebras-powered API service tier that ran GPT-5.6 Sol up to 14 times faster than the Standard tier. OpenAI described the preview in its announcement of Ultrafast mode.

What OpenAI announced

• Service tier: Ultrafast was an OpenAI API service tier, powered by Cerebras.
• Speed: in the preview, Ultrafast ran GPT-5.6 Sol up to 14x faster than the Standard tier and delivered up to 750 output tokens per second.
• Stage: the OpenAI API changelog described the launch as a limited preview.
• Expansion: Ultrafast expanded to GPT-6 Astra on September 29, 2026 and to GPT-6.1 Sol on October 8, 2026.

What it means for people building with AI

OpenAI offered speed as a separate API service tier, which matters for agent loops, where latency adds up over many model calls. The figures were stated as maximums: up to 14x faster than Standard for GPT-5.6 Sol and up to 750 output tokens per second, in a preview that the OpenAI API changelog called limited.

Availability and sources

The OpenAI API changelog entry lists Ultrafast as a limited preview for GPT-5.6 Sol. As a limited preview, it was not a general release at launch. The announcement also appeared in the OpenAI News RSS feed.

Sources

• Previewing Ultrafast mode: GPT-5.6 Sol at up to 14X the speed (OpenAI, August 13, 2026, official)
• OpenAI News RSS feed (item: Previewing Ultrafast mode) (OpenAI, August 13, 2026, official)
• OpenAI API changelog: Ultrafast limited preview for GPT-5.6 Sol (OpenAI, August 13, 2026, official)

Written with AI assistance, strictly from the sources listed above. Before it went live, automated checks verified its dates, sources and every language edition.