AWS puts OpenAI's GPT-6 Astra on Bedrock for business customers

AWS has made OpenAI's GPT-6 Astra generally available on Amazon Bedrock, adding OpenAI's latest frontier model to its managed AI platform for business customers.

Per IT Brief Asia, the model is available through supported Amazon Bedrock APIs, and customers can also configure ChatGPT Work and Codex to use GPT-6 Astra on Bedrock.

On capability, GPT-6 Astra supports a context window of up to 1 million input tokens, letting users submit large sets of material such as codebases, contracts and document collections. It is aimed at demanding business workflows involving reasoning, writing, design, and computer- or browser-based tasks.

Alongside the Bedrock release, OpenAI is introducing enterprise plugins for ChatGPT Work. According to AWS, those plugins extend Astra's browser-use functions across common business applications.

On governance and data, AWS said existing Bedrock controls can secure workloads, govern access and audit model invocation activity, and that inference data is not used for model training. For corporate buyers, those assurances matter as much as the model itself.

The release lands amid competition between cloud providers to host a growing range of third-party and in-house AI models for business customers. Putting another flagship OpenAI model on Bedrock widens the options for companies that want to build AI services without managing the underlying infrastructure.

The announcement came as part of AWS's latest weekly roundup of product updates. The same batch included Amazon Quick's desktop app becoming generally available for macOS and Windows, able to work with local files and stay connected to calendars, email and business applications in the background, with conversations, context and agents synced across desktop and mobile and agents continuing to run after a user closes their computer. AWS also raised the function timeout for asynchronous and event source mapping invocations on Lambda Managed Instances to 90 minutes from 15, while synchronous invocations keep the 15-minute maximum, and added cross-account copying and re-encryption for Amazon EBS Volume Clones using a KMS key in the target account.

The broader signal is the multi-cloud turn for frontier models: when the same flagship model appears both in its creator's own products and on major cloud platforms, enterprise choice shifts toward governance, cost and existing tooling rather than model access. The report did not specify the regions or pricing for GPT-6 Astra on Bedrock, and the exact set of business applications covered by the new ChatGPT Work plugins remains to be detailed — both worth tracking next.