AWS announced UltraFast mode availability for OpenAI’s GPT-6 Astra model on Amazon Bedrock. This premium speed tier is designed to serve workloads requiring rapid responses, delivering up to six times faster inference with speeds up to 300 tokens per second. It targets latency-sensitive scenarios like live coding assistants and interactive agents.
Amazon Bedrock provides a secure, reliable inference environment supporting control and governance features. The UltraFast mode helps applications requiring swift, high-quality outputs without compromising security or auditability. This mode is accessible via the Amazon Bedrock console and APIs, with region-specific endpoint availability detailed in official documentation.
Source & further reading
OpenAI GPT-6 Astra now supports UltraFast mode on Amazon Bedrock ↗Source published 2026-09-30. This report was prepared with AI assistance from the linked primary-source material. Claims and limitations should be checked against the original announcement or research.