Robert TaFounder & CEO, Clarity · Building Hyper-Personalized Agents

What Actually Changed in OpenAI's New Agents SDK

4 Jun 2026 · 1 min

Watch on YouTube

0:58

In this YouTube conversation, Robert Ta and Jonathan discuss the latest OpenAI Agents SDK update. Jonathan highlights that the SDK is now open‑source with Python implementations on GitHub and that OpenAI has partnered with Amazon AWS Bedrock, allowing the same agent harness to be deployed on Bedrock. They emphasize the simplification of the agent harness, noting Sam Altman's focus on "harness engineering" as crucial for model performance. A standout feature they mention is sandboxing, which lets developers create isolated environments to test and run agents safely. Overall, the discussion frames the new SDK as a more accessible, deployable, and developer‑friendly toolkit for AI builders.

Chapters

You'll see here we have OpenAI, the next evolution of the Agents SDK. Jonathan, I guess you and I have talked about this in our team meetings, but any big things that stand out to you with the new Agents SDK from an AI builder perspective?
Yeah, so OpenAI released this update a couple weeks ago and it was an evolution of their SDK. I think one thing, it's open source, so there's Python versions available on GitHub, which is cool. And the other thing is that they're announcing a partnership with Amazon AWS Bedrock. So you can use the same harness to deploy your agents on Bedrock. So that's a pretty empowering set of capabilities. The first thing I took away is just like how simplified the agent harness was. And you've heard like the word harness, harness engineering has come up a lot. Sam Altman himself is saying like how critical building the harness is to getting the models to work and how it's just kind of embedded now. A lot of great tools they provided here. I think maybe the first one I was drawn to is the sandboxing capability and creating environments for your agent that you can run. And test them out in, in a sandbox environment.