ChatGPT, OpenAI's text-generating AI chatbot, has taken the world by storm since its launch in November 2022. What started as a tool to supercharge productivity through writing essays and code with short text prompts has evolved into a behemoth with 300 million weekly active users by early 2025, and later surpassing 800 million weekly users by October. In 2025, OpenAI battled perception it was losing ground to Chinese rivals like DeepSeek, while shoring up relations with Washington, pursuing ambitious data center projects, and laying groundwork for one of the largest funding rounds in history. The year also saw internal drama, high-profile executive exits, and lawsuits over copyright and mental health risks. Below is a timeline of the most significant ChatGPT updates in 2025.

January 2025

o3-mini Reasoning Model

OpenAI launched o3-mini, a new AI reasoning model designed to be both powerful and affordable. The model showcased improved chain-of-thought reasoning, with more visible step-by-step processing to address user demands for transparency.

Operator AI Agent Preview

OpenAI introduced Operator, a general-purpose AI agent that can control a web browser to autonomously perform tasks like booking travel, making reservations, and shopping online. It was first made available as a research preview for Pro subscribers at $200/month.

New Customization Features

ChatGPT gained the ability to schedule reminders and recurring tasks via a new "tasks" beta feature. Users could also assign personality traits like "Chatty" or "Gen Z" to the chatbot, enhancing personalization.

February 2025

GPT-5 Announcement and o3 Cancellation

OpenAI cancelled the standalone release of o3 in favor of a unified GPT-5 model that integrates multiple technologies, including o3. CEO Sam Altman described this as a simplified product offering.

Deep Research Agent

OpenAI unveiled deep research, an AI agent designed for in-depth, complex research tasks that require synthesizing information from multiple sources.

ChatGPT Web Search Without Login

Anyone could now use ChatGPT web search without logging in, though responses were limited to the chatbot's last training update unless signed in.

March 2025

Image Generation Upgrades and Ghibli Trend

OpenAI rolled out a major upgrade to ChatGPT's image generation using GPT-4o, enabling creation of Studio Ghibli-style images. The feature went viral but raised copyright concerns. OpenAI later updated policies to allow generation of public figures and hateful symbols.

New Models and API Updates

OpenAI released new transcription and voice-generating AI models, including gpt-4o-mini-tts and improved speech-to-text models. The company also launched tools for businesses to build AI agents via the Responses API.

Leadership Shifts

Brad Lightcap took on expanded responsibilities as COO leading global expansion, while Sam Altman focused on research and products. Mark Chen became chief research officer.

April 2025

GPT-4.1 Family and o3/o4-mini

OpenAI released GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano focused on coding, alongside reasoning models o3 and o4-mini. The o3 model showed high hallucination rates but advanced reasoning capabilities.

Sycophancy Fixes

After an update made ChatGPT overly flattering and agreeable, OpenAI rolled back the change and worked on additional fixes to prevent sycophancy.

Social Media Network Plans

OpenAI began early development of its own social media platform to compete with X and Threads, potentially integrated into ChatGPT.

May 2025

Codex AI Coding Agent

OpenAI introduced Codex, a specialized AI coding agent powered by codex-1, capable of fixing bugs, writing features, and running tests over periods from one minute to over seven hours.

Data Residency in Asia

OpenAI launched a data residency program for ChatGPT Enterprise and Edu in India, Japan, Singapore, and South Korea, allowing local data storage.

Hardware Acquisition

OpenAI agreed to purchase Jony Ive's devices startup IO for $6.4 billion, aiming to enhance ChatGPT with hardware.

June 2025

Voice Mode Upgrade

Advanced Voice Mode was upgraded for all paid users, enabling more natural and fluid conversations with ChatGPT using improved speech synthesis.

Google AI Chips

OpenAI started using Google's AI chips (TPUs) to power ChatGPT and other products, marking a shift from exclusive reliance on Nvidia GPUs.

MIT Study on Critical Thinking

A study from MIT's Media Lab suggested that ChatGPT use may harm critical thinking skills, as users showed minimal brain engagement when writing essays.

July 2025

ChatGPT Agent Launch

OpenAI launched ChatGPT Agent, a general-purpose AI agent capable of navigating calendars, drafting presentations, shopping online, and handling complex workflows within a secure virtual environment.

Study Mode

OpenAI introduced Study Mode, a feature designed to promote critical thinking by prompting students to engage with material rather than providing direct answers.

Delayed Open Model

OpenAI again delayed its open-weight model launch due to additional safety testing, with no set release date.

August 2025

GPT-5 Launch

OpenAI released GPT-5, a next-gen AI model that automatically chooses between Auto, Fast, and Thinking modes. It offered improved task handling and reasoning, with rate limits of 3,000 messages per week for Thinking mode.

Open Source Models

OpenAI released two open-weight language models: gpt-oss-120b and gpt-oss-20b, marking a return to open-source AI after years of proprietary releases.

Federal Government Push

OpenAI offered ChatGPT Enterprise to U.S. federal agencies for just $1 for the first year, aiming to secure government contracts.

September 2025

Parental Controls

OpenAI added parental controls to ChatGPT, allowing parents to link accounts, set quiet hours, limit sensitive content, and disable features like voice mode.

Pulse Morning Briefs

OpenAI unveiled Pulse, a feature delivering personalized morning briefings overnight, starting with Pro users.

Shopping Integration

Instant Checkout was launched in ChatGPT for U.S. users, enabling direct purchases from Etsy and Shopify merchants without leaving the conversation.

October 2025

Atlas AI Browser

OpenAI launched ChatGPT Atlas, an AI-powered browser for Mac (with Windows, iOS, Android coming soon) that integrates ChatGPT as the primary search tool.

Business Platform Milestone

OpenAI announced over 1 million businesses using its products, making it the fastest-growing business platform in history.

Mental Health Conversations

OpenAI revealed that ChatGPT handles over a million suicide-related conversations weekly, and improved responses by consulting mental health experts.

November 2025

GPT-5.1 and GPT-5.2

OpenAI released GPT-5.1 with warmer conversational tone and faster simple-task handling, followed by GPT-5.2 in three versions (Instant, Thinking, Pro) to compete with Google.

Group Chats

ChatGPT group chats became available to all users, allowing collaboration in shared conversations.

Legal Challenges

Seven families sued OpenAI over ChatGPT-related suicides, and a Munich court ruled that ChatGPT violated German copyright law by reproducing song lyrics.

December 2025

Energy and Tone Controls

OpenAI added new controls for adjusting ChatGPT's warmth, enthusiasm, emoji use, and formatting style to address sycophancy concerns.

Disney Partnership

Disney invested $1 billion in OpenAI, bringing over 200 characters to Sora video generation for exclusive use in the first year.

Code Red Memo

CEO Sam Altman declared a "code red" internally, prioritizing ChatGPT improvements over other initiatives like advertising as competition from Google intensified.

GPT Image 1.5

OpenAI released GPT Image 1.5, a faster and more precise image generation model to compete with Google Gemini and other rivals.

Source: TechCrunch News