AWS Weekly Roundup: GPT-6.1 Terra on Bedrock, Step Functions as agent tools, and more (November 2, 2026)

Summary

• OpenAI GPT-6.1 Terra is generally available on Amazon Bedrock, scoring 15 percentage points higher than GPT-5.6 Terra in OpenAI's Agents' Last Exam evaluation while performing on par with GPT-6 Astra.
• GPT-6.1 Terra completes tasks more than three times faster than GPT-6.1 Sol and GPT-5.6 Terra, with input and output token prices at least 50% lower than those models.
• GPT-6.1 Terra supports up to 1 million input tokens with text and image input via Bedrock APIs, ChatGPT Work, and Codex.
• AWS Step Functions workflows are now available as AI agent tools through Amazon Bedrock AgentCore Gateway using the Model Context Protocol.
• AWS Key Management Service now supports ML-KEM for post-quantum key encapsulation with private keys in FIPS 140-3 Level 3 validated HSMs.
• AWS KMS supports all three NIST FIPS 203 parameter sets for Encapsulate and Decapsulate operations, excluding Middle East (UAE, Bahrain) and China Regions.
• Amazon Aurora PostgreSQL versions 18 and higher added post-quantum TLS key exchange.
• AWS IAM Identity Center now supports cross-Region trusted identity propagation, allowing user identities to travel with requests to services in other Regions.
• Amazon ECS now supports GPU sharing with NVIDIA Multi-Process Service to run multiple tasks concurrently on a single GPU without code changes.
• AWS Lambda supports Seekable OCI lazy loading for Lambda Managed Instances, reducing container startup time by up to 70%.
• Amazon SageMaker AI training and processing jobs also support Seekable OCI lazy loading, with benchmarks showing 3 to 11 seconds startup time regardless of image size.
• Amazon EKS Auto Mode now supports Kubernetes Gateway API for traffic declaration without installing a separate controller.
• Nubank used the AWS DevOps Agent to reduce a test cycle from 4.3 hours to roughly 40 minutes and quadrupled flow coverage to more than 160 in 2026.
• MoEngage exposed roughly 400 internal REST API operations as agent tools using AgentCore Gateway, saving an estimated six months of platform engineering.
• A case study on S3 storage costs showed that transitioning 5.8 PB of logs to Glacier Instant Retrieval cut the monthly bill by roughly 70%.
• AWS re:Invent returns to Las Vegas from November 30 to December 4 with more than 2,200 sessions.
• AWS Community Days 2026 season is winding down as re:Invent approaches.