OpenAI Launches New Tools For Developers & Enterprises To Build Advanced AI Agents

Recent reports suggested that generative artificial intelligence (GenAI) industry leader OpenAI was planning to launch specialized AI agents tailored for different applications and charge businesses to use them—with prices ranging from $2,000 for “high-income knowledge worker” agents, to $20,000 per month for an agent capable of “PhD-level research.”

Now, OpenAI has published a blog post announcing the launch of new tools—which it called the first set of building blocks—that will help developers and enterprises build useful and reliable agents.

Agents are designed to execute complex real-world tasks independently (without relying on human intervention). They can be built to address specific processes or operations in diverse industries while providing focused and dedicated outputs.

For OpenAI, agents are systems that can accomplish tasks on behalf of users, and this is what the company wants to enable businesses and developers to do.

While the previous year has seen the company introduce a wide range of new model capabilities, including advanced reasoning, multimodal interactions, new safety techniques, and more, it’s noted that users face difficulties in transforming such capabilities into production-ready agents. Challenges include requiring “extensive prompt iteration and custom orchestration logic without sufficient visibility or built-in support.”


To curb these problems, OpenAI is launching a new set of APIs and built-in tools specifically designed to simplify the development of agentic applications. The new tools are meant to streamline core agent logic, orchestration, and interactions, making it easier for developers to commence building agents. The new tools include:

Responses API

Responses API forms OpenAI’s new API primitive for leveraging OpenAI’s built-in tools to build agents, combining the simplicity of Chat Completions with the tool-use capabilities of the Assistants API.

Responses API brings with it several usability improvements that will enable it to support new built-in tools such as web search, file search, and computer use, and will work together to connect models to the real world.

It’s designed to allow developers to easily combine OpenAI models and built-in tools into their apps, without having to deal with multiple APIs or external vendors, allowing them to solve increasingly complex tasks with a single Responses API call. Furthermore, it makes it easier to store data on OpenAI (which will not be used by OpenAI to train its models).

Responses API is available now and is not charged separately but instead billed at standard rates.

Built-In Tools Within Responses API

OpenAI is bringing a variety of built-in tools in the Responses API to make it easier for developers to build AI agents.

Web Search will bring developers updated answers quickly and with clear and relevant sources from across the web. It will be available as a tool through the gpt-4o and gpt-4o-mini models and can be paired with other tools or function calls.

Web Search is powered by the same model used for ChatGPT search and is available to all in preview in Responses API, along with direct access to OpenAI's fine-tuned search models in the Chat Completions API. Pricing starts at $30 and $25 per thousand queries for GPT‑4o search and 4o-mini search respectively.

File Search⁠ allows developers to easily retrieve relevant information from large volumes of documents. It comes with support for multiple file types, query optimization, metadata filtering, and custom reranking, and can deliver fast, accurate search results. Using Responses API, developers can integrate it with only a few lines of code.

This capability can be used for multiple purposes, including “enabling a customer support agent to easily access FAQs, helping a legal assistant to quickly reference past cases for a qualified professional, and assisting a coding agent to query technical documentation.” Available now, it’s priced⁠ at $2.50 per thousand queries and file storage at $0.10/GB/day, with the first GB free.

Computer Use enables developers to build agents capable of completing tasks on a computer through the Responses API. It’s powered by the same Computer-Using Agent (CUA) model that enables Operator. The tool captures mouse and keyboard actions generated by the model, which allows developers to automate computer use tasks by translating such actions into executable commands. Additionally, it can be used to automate browser-based workflows.

Available now as a research preview for select developers in usage tiers 3-5⁠, Computer Use is priced at $3/1M input tokens and $12/1M output tokens.

Agents SDK

OpenAI is offering developers its new open-source Agents SDK to simplify orchestrating multi-agent workflows, a move that goes beyond just building the core logic of agents and providing them with access to tools.

Agents SDK comes with improvements over OpenAI’s previously launched experimental SDK—Swarm—and can help orchestrate single-agent and multi-agent workflows.

As such, the improvements include easily configurable LLMs with clear instructions and built-in tools, enhanced transfer control between agents, better configurable safety checks, and stronger tracing and observability to debug and optimize performance.

Agents SDK works with the Responses API and Chat Completions API, and can be used for numerous real-world applications, including customer support automation, multi-step research, content generation, code review, and sales prospecting.

Integrated Observability Tools⁠

OpenAI’s integrated observability tools⁠ are meant to trace and inspect agent workflow execution, ensuring developers effectively build, deploy, monitor, evaluate, and improve their agentic systems. This allows them to observe agent behavior in real-time and debug issues, measure agent performance, identify areas for improvement, and refine agents.


The new move will see Responses API replace Assistants API, which is scheduled to be phased out by the second half of 2026.

OpenAI’s move comes as Chinese companies DeepSeek and the more recent Manus challenge its capabilities in the artificial intelligence (AI) industry.

Do you think this move will help OpenAI thwart threats coming from Chinese AI companies?

Let us know in the comments below!