AWS Weekly Roundup: Price Cuts for GPT Models on Bedrock, CloudWatch Managed Collector for Prometheus Metrics, and More (August 3, 2026)

Amazon Bedrock has reduced on-demand inference pricing for OpenAI's GPT-5.6 models. Starting July 30, GPT-5.6 Luna pricing dropped 80% to $0.20 per 1 million input tokens and $1.20 per 1 million output tokens, while GPT-5.6 Terra pricing dropped 20%. The pricing changes apply automatically with no action required.

Amazon CloudWatch added a managed collector for gathering Prometheus metrics from AWS infrastructure. It enables monitoring of Amazon EKS, EC2, ECS, MSK, and OpenSearch Service workloads without deploying or managing separate agents.

AWS also announced the following updates:

• Multicloud connectivity between AWS Interconnect and Oracle Cloud Infrastructure is now generally available. It enables scalable, resilient private connections between AWS and OCI without traversing the public internet.
• Multi-Region support for IAM Identity Center directories has been expanded. If the home Region experiences a disruption, users can continue accessing AWS accounts using permissions provisioned in additional Regions.
• Amazon S3 Tables now supports the Variant data type from Apache Iceberg V3. This allows semi-structured data with flexible schemas, such as IoT sensor data and application logs, to be handled without JSON blobs.
• AWS also published a method for installing and updating the AWS CLI with a single command across multiple platforms.