AWS announced that GPT‑6 Astra from OpenAI is now available with UltraFast mode on Amazon Bedrock. The new tier is aimed at workloads where speed is critical, such as real‑time coding assistants and interactive agents. AWS says the mode can run up to six times faster than the standard API, delivering about 300 tokens per second. Users can enable it through the Bedrock console or its APIs, subject to regional availability and pricing.
Why it matters
Developers can build latency‑sensitive AI features, like live code helpers, that respond noticeably quicker.
