All news with #amazon bedrock tag

🔒 BeyondTrust researchers demonstrated that the Sandbox mode in AWS Bedrock AgentCore Code Interpreter permits outbound DNS A/AAAA queries that can be abused to create a bidirectional covert channel. By encoding data in DNS requests and responses they showed both data exfiltration and an interactive reverse shell without triggering network restrictions. AWS reproduced the report but characterized the behavior as intended and updated documentation rather than issuing a patch.

🚀 Amazon Web Services has launched Amazon Bedrock in the Asia Pacific (New Zealand) Region, enabling customers to build and scale generative AI applications using a single API and a choice of foundation models. The managed service emphasizes built-in security, privacy, and responsible AI capabilities to support enterprise deployments. Models now available in New Zealand include Anthropic (Sonnet 4.5, 4.6; Opus 4.5, 4.6; Haiku 4.5) and Amazon’s Nova 2 Lite with cross-region inference support.

🤖 AWS announces general availability of AWS Partner Central agents, AI-powered agentic capabilities built on Amazon Bedrock AgentCore to accelerate partner co-selling. Agents provide pipeline insights, tailored sales plays, and next-step recommendations, and can populate CRM fields from transcripts, notes, and emails. They also identify funding eligibility, pre-fill funding requests, and are available in all commercial AWS Regions.

⚠️ Phantom Labs Research demonstrated a DNS-based exfiltration technique targeting the AWS Bedrock AgentCore Code Interpreter that bypasses expected Sandbox Mode network restrictions. Maliciously crafted files (for example, CSVs) can influence generated Python code to use DNS queries as a covert command-and-control channel. In tests, researchers executed commands, enumerated and retrieved S3 content and secrets while the environment still reported network access disabled. AWS says this is intended behavior and updated documentation; organisations should inventory AgentCore instances, tighten IAM roles and move sensitive workloads to VPC mode.

⚙️ Amazon Bedrock AgentCore Runtime now supports the Agent-User Interaction (AG-UI) protocol, enabling developers to deploy AG-UI servers that deliver real-time, interactive agent experiences into user-facing applications. AgentCore Runtime handles authentication, session isolation, and autoscaling for AG-UI workloads so teams can focus on building responsive frontends. AG-UI streams text, reasoning steps, and tool results over SSE and WebSocket and is available across fourteen AWS Regions.

🔔 Amazon announced that AgentCore Memory for Bedrock now supports streaming notifications for long-term memory, delivering push events to Amazon Kinesis whenever memory records are created or modified. Developers can subscribe to these streams to trigger downstream workflows, refresh application state, and build audit trails without polling. This reduces developer overhead and simplifies integration of personalized, memory-driven experiences. The feature is available in 15 AWS Regions.

🔔 Amazon Web Services has added stateful Model Context Protocol (MCP) server capabilities to Amazon Bedrock AgentCore Runtime, enabling server-driven elicitation, sampling, and progress notifications alongside existing support for resources, prompts, and tools. Developers can now run each session in an isolated microVM and preserve session context across interactions using an Mcp-Session-Id header. These features support interactive, multi-turn workflows and real-time updates for long-running operations.

🔍 Amazon Bedrock now emits two CloudWatch metrics: TimeToFirstToken and EstimatedTPMQuotaUsage. TimeToFirstToken measures latency from request submission to the first returned token for streaming APIs, enabling SLA baselines and latency alarms without client-side instrumentation. EstimatedTPMQuotaUsage estimates Tokens Per Minute consumption across Bedrock inference APIs to support proactive quota alarms and capacity planning.

🔍 AWS Elastic Beanstalk now offers AI-powered environment analysis that collects recent events, instance health, and logs and sends them to Amazon Bedrock for automated diagnosis. Developers and operations teams can request an analysis from the Elastic Beanstalk console or use the AWS CLI with RequestEnvironmentInfo and RetrieveEnvironmentInfo. The feature returns step-by-step, environment-specific troubleshooting recommendations to help reduce mean time to resolution and is available in Regions where both services exist.

🤖 Amazon Lightsail now lets you deploy OpenClaw, a private self-hosted AI assistant, on your own cloud infrastructure with simple, secure defaults. Each Lightsail OpenClaw instance includes built-in security controls—sandboxed agent sessions, one-click HTTPS for TLS, device-pairing authentication, and automatic snapshots—reducing manual configuration and operational risk. Amazon Bedrock is the default model provider, and users can swap models or connect to Slack, Telegram, WhatsApp, and Discord as needed.

🔒 Amazon has made Policy in Bedrock AgentCore generally available, providing centralized, fine-grained controls for agent-to-tool interactions. Teams can author policies in natural language that AWS converts into Cedar and stores in a policy engine attached to an AgentCore Gateway, which intercepts traffic and evaluates requests before allowing or denying access. Operating outside agent code, this lets security, compliance, and operations enforce access rules and validate inputs without modifying agents, improving governance and visibility across deployments.

🔁 Amazon Bedrock batch inference now accepts the Converse API as a model invocation type, letting you submit batch inputs in a consistent, model-agnostic Converse request format and receive outputs in the Converse response format. This unifies real-time and batch request formats, simplifying prompt management and reducing the effort of switching between models. You can configure the Converse invocation type through the Bedrock console or API, and the capability is available in all Regions that support Bedrock batch inference.

🚀 Amazon Bedrock now offers an OpenAI-compatible Projects API within the Mantle inference engine, enabling customers to create isolated projects for separate applications, environments, or teams. Each project supports distinct IAM-based access controls and tagging to improve security boundaries and cost visibility. The feature is available for OpenAI-compatible APIs, the Responses API, and Chat Completions through Mantle. There is no additional charge beyond model inference consumption.

🔒 In November 2025, AWS completed its first surveillance audit for ISO/IEC 42001:2023 — the Artificial Intelligence Management System standard — with no findings. This follows AWS’s November 2024 announcement that several AI services, including Amazon Bedrock, Amazon Q Business, Amazon Textract, and Amazon Transcribe, were accredited under the standard. The successful no-findings outcome provides independent validation of AWS’s ongoing commitment to responsible AI practices and gives customers added assurance when building and operating AI applications on AWS.

🔧 Amazon Bedrock now supports server-side tool execution by integrating AgentCore Gateway with the Responses API. Customers can supply an AgentCore Gateway ARN and Bedrock will discover and invoke gateway tools during inference, eliminating client-side orchestration. Tool results are executed and streamed server-side in real time, with IAM-based access control preserved. This reduces latency and simplifies agentic workflows.

🔎 AWS added source document references to Automated Reasoning policies in Amazon Bedrock, enabling users to trace generated formal rules and variables back to the original documents used to create policies. This improves transparency and makes it easier to review, validate, and refine policy encodings derived from uploaded materials. Test generation for Automated Reasoning checks is now available in multiple US and Europe Regions and can be accessed via the Bedrock console and the Bedrock Python SDK. AWS says the checks use formal verification and can reach up to 99% accuracy in identifying correct LLM responses, helping detect hallucinations and ambiguity.

🚀 Amazon Bedrock now offers Mistral AI Devstral 2 123B, an open-weight 123B-parameter LLM optimized for agentic software engineering workflows. The model focuses on code generation, automation, and reliable multi-step reasoning, supporting long-context comprehension for multi-turn coding tasks. Bedrock exposes Devstral 2 via a single, fully managed API so customers do not need to provision infrastructure or host models. It is intended for production coding assistants, automated code review, and complex software development agents and is available in select AWS Regions.

🔧 Amazon Bedrock now supports reinforcement fine-tuning (RFT) for open-weight models, including openai.gpt-oss-20b and qwen.qwen3-32b. The managed RFT workflow automates end-to-end customization using reward functions that can be rule-based or AI-driven, and integrates with AWS Lambda for custom grading and checkpoint inspection. Fine-tuned models are immediately available for on-demand inference via Bedrock's OpenAI-compatible Responses and Chat Completions APIs, while proprietary data remains within AWS's secure environment.

🚀 Amazon Bedrock now supports Claude Sonnet 4.6, Anthropic's newest model delivering frontier performance across coding, agentic workflows, and professional knowledge work. Sonnet 4.6 enables faster, high-quality task completion and claims near-human reliability for browser-based automation, at a lower cost than Opus 4.6. Enterprises can migrate from Sonnet 4.5 with minimal prompt changes and deploy it across supported Bedrock regions for search, chat, agents, and domain-specific applications.

🚀 Amazon Web Services announced that Amazon Bedrock now supports the latest open-weight models in Asia Pacific (Sydney) through the bedrock-mantle endpoint. The update brings models from providers including DeepSeek, Google, MiniMax, Mistral, Moonshot AI, Nvidia, and OpenAI, expanding local model choice. Powered by Project Mantle, bedrock-mantle delivers a distributed, serverless inference engine with advanced quality-of-service controls, automated capacity management and unified pools. It also offers out-of-the-box OpenAI API compatibility to simplify integration for developers.