OpenAI upgrades Responses API with features built specifically for long-running AI agents

2026-02-11 · Source: The Decoder · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Intermediate, quick

Summary

OpenAI has enhanced its Responses API with new features designed for long-running AI agents. The update introduces server-side compression, enabling agent sessions to persist for hours without exceeding context limits. Additionally, OpenAI-hosted containers now have controlled internet access, allowing them to install libraries and execute scripts. A key new feature is "skills," which are reusable, versioned bundles of instructions, scripts, and files that agents can access and run on demand. These skills function as an intermediary layer between system prompts and tools, allowing developers to package complex workflows into modular ZIP files that activate only when required, supporting both hosted and local container environments.

Key takeaway

For AI Architects designing persistent, complex agent systems, OpenAI's updated Responses API offers critical capabilities. Server-side compression and controlled internet access directly address common operational hurdles, while the new "skills" feature provides a robust framework for modularizing agent behaviors. You should evaluate integrating these features to enhance agent longevity, reduce context window strain, and streamline workflow management in your deployments.

Key insights

OpenAI's API now supports long-running AI agents via server-side compression, internet access, and modular "skills."

Principles

• Modularize complex agent workflows.
• Pin specific skill versions in production.

Method

Package agent workflows as versioned ZIP files ("skills") for on-demand execution, functioning as a middle layer between system prompts and tools.

In practice

• Use server-side compression for extended agent sessions.
• Enable internet access for library installation.
• Build skills like small command-line programs.

Topics

• OpenAI API
• AI Agents
• Agent Skills
• Context Management
• API Features

Best for: AI Architect, CTO, VP of Engineering/Data, AI Engineer, Machine Learning Engineer, Software Engineer

Related on AIssential

• Codex Subagents, Minimax M2.7, Claude Code Update, NVIDIA GTC, AI CMO, Google AI Tool! HUGE AI News — AI agents are evolving towards greater autonomy, specialization, and local execution across development, design, and operational tasks.
• A Primer on Using Agent Skills — Agent skills provide a modular, dynamic way to equip AI agents with specific capabilities, improving efficiency and reliability.
• 🔵 Slack gets 30 New Features at Once — AI-driven development accelerates feature velocity, transforming platforms into agentic operating systems and necessitating new skill frameworks.
• The OpenClaw-ification of AI — AI products are evolving towards autonomous, persistent, and context-aware agentic systems, mirroring OpenClaw's capabilities.
• TAI #191: Opus 4.6 and Codex 5.3 Ship Minutes Apart as the Long-Horizon Agent Race Goes Vertical — AI agent capabilities are rapidly advancing, with new models demonstrating significant leaps in multi-step task endurance and context handling.
• xAI Releases Grok Skills and Updates Tool Calling Responses API — Grok Skills provides persistent, customizable AI expertise and integrated tool calling for automated workflows across platforms.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Decoder.