Amazon Web Services has made OpenAI’s GPT-6 Astra generally available on Amazon Bedrock. Enterprises can call the model through supported Bedrock APIs or configure ChatGPT Work and Codex to route requests to Astra running on the Bedrock inference engine.
The model supports a context window of up to 1 million input tokens and is positioned for complex reasoning, coding, computer use, research, and document creation. OpenAI is also introducing new enterprise plugins for ChatGPT Work that extend Astra’s browser-use capabilities across common business applications.
Key takeaways
- GPT-6 Astra is now generally available on Amazon Bedrock for business customers.
- Context window reaches approximately 1 million input tokens; max output listed at 128,000 tokens.
- ChatGPT Work and Codex can be configured to use Astra on Bedrock; new enterprise plugins ship alongside the launch.
What OpenAI’s Astra threshold actually means on Bedrock
AWS positions Astra as the latest and most capable OpenAI model available on the platform. Customers keep inference data inside AWS controls; the announcement states that using GPT-6 Astra does not require opting into sharing data with OpenAI for training. Model IDs and endpoint details appear in the Bedrock documentation under openai.gpt-6-astra.
Primary sources are the AWS what’s-new post and the corresponding Bedrock model card. For related creator-stack context see How AI is Reshaping the Content Creator Stack in 2026.
What to watch next
Watch whether other frontier models follow the same Bedrock path and how enterprise pricing tiers settle once volume builds. The Critical cybersecurity classification noted for Astra will also shape how regulated customers adopt the model for agent workloads.