OpenAI models on Amazon Bedrock from Australia

This launch extends the availability of OpenAI's GPT-5.6 models to Australian teams through Amazon Bedrock. The use of global cross-Region inference allows applications in Sydney and Melbourne to route requests to a supported commercial AWS Region, providing a larger capacity pool without extra management overhead. This is relevant for businesses that need to scale AI workloads while maintaining low latency. The models are described for different use cases: Sol for demanding reasoning and coding tasks, Terra for everyday production use, and Luna for high-volume, latency-sensitive applications. They all support 1 million token context windows, which could help with processing large documents or long conversations. AWS also provides tools for cost optimization (prompt caching), integration with Codex, and monitoring via CloudWatch, which could assist teams in managing their AI operations.