The next evolution of the Agents SDK
OpenAI has updated its Agents SDK with native sandbox execution and a model-native harness, enabling developers to build secure, long-running agents that operate across files and tools. The update targets production-grade agentic workflows by providing safer code execution environments and tighter integration with OpenAI models. This represents a continued push by OpenAI to mature its developer tooling for autonomous agent deployment.
Related guides (4)
Related events (8)
New Tools for Building Agents
OpenAI announced new tools aimed at developers building AI agents, published on March 11, 2025. The announcement comes from OpenAI's official blog, signaling a continued push to expand the agent-building ecosystem. Specific tools and capabilities were not detailed in the provided body text, but the source and framing indicate a product/tooling release targeting the agentic development workflow.
OpenAI Introduces AgentKit, Expanded Evals, and Reinforcement Fine-Tuning for Agents
OpenAI has released a suite of developer tools aimed at accelerating agent development from prototype to production. The release includes AgentKit (a new agent-building framework), expanded evaluation capabilities, and reinforcement fine-tuning (RFT) specifically designed for agentic use cases. These tools represent OpenAI's continued push to provide end-to-end infrastructure for building and deploying AI agents at scale.
From model to agent: Equipping the Responses API with a computer environment
OpenAI describes how it built an agent runtime by combining the Responses API with a shell tool and hosted containers, enabling agents to operate with persistent files, tools, and state. The architecture supports secure, scalable execution of agentic workflows. This represents a concrete infrastructure layer for deploying agents in production environments.
Running Codex Safely at OpenAI
OpenAI published a blog post describing the security architecture used to run Codex as a coding agent internally, covering sandboxing, human approval workflows, network policies, and agent-native telemetry. The post is aimed at supporting enterprise adoption of coding agents by demonstrating safe and compliant deployment patterns. It provides operational detail on how OpenAI itself governs agentic code execution in production.
Introducing the Stateful Runtime Environment for Agents in Amazon Bedrock
OpenAI and Amazon Web Services are launching a Stateful Runtime Environment for Agents in Amazon Bedrock, enabling persistent orchestration, memory, and secure execution for multi-step AI agent workflows. The integration brings OpenAI's models into AWS's managed agent infrastructure with stateful capabilities. This represents a significant enterprise deployment partnership between two major AI ecosystem players.
Agent-S: Open Agentic Framework for Human-Like Computer Use
Agent-S is an open-source Python framework by Simular AI designed to enable AI agents to interact with computers in a human-like manner. The project has accumulated 11,388 GitHub stars with modest daily growth of 29 stars. It represents an entry in the growing space of computer-use agent frameworks targeting GUI and desktop automation tasks.
Microsoft agent-framework: open-source library for building and orchestrating AI agents
Microsoft has published an open-source framework on GitHub for building, orchestrating, and deploying AI agents and multi-agent workflows, with support for both Python and .NET. The repository has accumulated 11,061 stars. It represents Microsoft's entry into the agent harness tooling space alongside existing frameworks like LangChain and AutoGen.
License to Call: Introducing Transformers Agents 2.0
Hugging Face announced Transformers Agents 2.0, a major update to their agent framework built on top of the Transformers library. The release introduces new abstractions for tool use, multi-step reasoning, and agent orchestration, positioning it as a production-ready framework for building AI agents. The update reflects growing ecosystem investment in standardized agent tooling patterns.