🐺 OpenAI Dev Day: Game-Changing Releases

Hi there fellow AI enthusiast,

This week’s newsletter brings:

OpenAI Dev Day 2025

This week, OpenAI introduced the next generation of tools and models to help developers code faster, build more reliable agents, and scale their apps directly inside ChatGPT.

Apps in ChatGPT

You can now chat with apps directly in ChatGPT and build them using the Apps SDK (preview).
Example: “Spotify, make a playlist for my party this Friday.” ChatGPT recognizes the app, brings it into your conversation, and asks for permission the first time you connect.
📖 Read more: OpenAI – Apps SDK

AgentKit

AgentKit is a complete toolkit for building, testing, and deploying production-grade agents.
Key components:

• Agent Builder → a visual canvas for creating and versioning multi-agent workflows.

• Connector Registry → a central place to manage data and tool connections.

• ChatKit → a toolkit to embed chat-based agent experiences into your product.
  📖 Read more: OpenAI – AgentKit

Sora 2 in the API

The new Sora 2 model lets you integrate video generation directly into your app.

from openai import OpenAI
openai = OpenAI()

video = openai.videos.create(
    model="sora-2",
    prompt="A video of a cool cat on a motorcycle in the night",
)

📖 Read more: OpenAI – Video Generation Guide

Codex, Now Generally Available

Codex is officially out of preview, with new features like a Slack integration, SDK, and enterprise controls.
You can now use it across your editor, terminal, and the cloud — all linked to your ChatGPT account.
Daily usage has grown more than 10x since August, and GPT-5 Codex is already serving over 40 trillion tokens in just three weeks.
📖 Read more: OpenAI – Codex

GPT-5 Pro in the API

OpenAI’s most advanced and precise model yet, built for tasks where accuracy really matters. It is the most expensive model to date.
💰 Input: $15.00 | Output: $120.00 per 1M tokens.
📖 Read more: OpenAI – GPT-5 Pro

gpt-realtime-mini & gpt-image-1-mini

Two new lightweight and cost-efficient models:

• gpt-realtime-mini → real-time voice model, 70% cheaper than the large version.

• gpt-image-1-mini → image generation model, 80% cheaper.

📖 Read more: Realtime Mini | Image Mini

Thank you and see you next time!

PS: If you feel your team would benefit from knowing how to apply the AI tools and features covered in this email, I’d love to help. All you need to do is hit the reply button for us to get started. (Free consultation, no strings attached).