AWS Transform continuous modernization is now live across all supported regions. This tool helps your engineering team tackle technical debt by analyzing source code repositories at scale—connect GitHub, GitLab, or Bitbucket, run analyses on demand, and get automated pull requests with validated code fixes. Work from the web app, your IDE, or terminal while keeping full control of your code.
OpenAI's GPT-5.6 Sol, Terra, and Luna now support massive 1 million token context windows on Amazon Bedrock, letting you process entire codebases and lengthy documents without chunking. The models deliver more accurate responses while prompt caching keeps costs down for repeated context. Available across multiple AWS regions with easy setup via Bedrock Console or API.
Amazon GameLift Streams now lets you share playable game sessions via simple URLs—no AWS account or software needed. Just create a stream URL, set expiration and session limits, and share the link. Each recipient gets their own independent stream session routed to a nearby location. You can manage everything through the console or new APIs, with no extra charges beyond standard streaming capacity.
AWS Resilience Hub now recommends tailored resilience tests to help your teams validate service recovery from real failure scenarios. Using AWS Fault Injection Service, it automatically injects controlled faults and checks if your systems recover within your targets. Tests cover Availability Zone failures, regional outages, and dependency issues—perfect for platform and SRE teams.
AWS Organizations users can now check their account quota and usage directly in Service Quotas—no need to contact support anymore. This makes it easier to plan your account growth and request increases before hitting limits. The feature is available now in US East (N. Virginia).
AWS Lambda now supports up to 10,000 event pollers for Amazon SQS (5x increase from 2,000), enabling 100,000 concurrent invocations per mapping. Perfect for mission-critical apps like real-time order processing and IoT telemetry. Provisioned Mode lets you optimize throughput without splitting workloads across multiple mappings.