ChatGPT 2025 Updates: Features, Growth, and OpenAI’s Challenges

ChatGPT 2025 updates have transformed OpenAI’s flagship product into a multi-modal AI powerhouse with 300 million weekly users and a flood of new features. From voice interaction to deep research tools, ChatGPT has evolved rapidly — but not without controversy.

OpenAI’s journey through 2025 has been marked by innovation, lawsuits, executive exits, and increased global scrutiny. Below is a comprehensive timeline of the year’s highlights.

June 2025: Hardware Shift, Brain Studies & Downloads Surge

OpenAI began using Google’s AI chips alongside Nvidia GPUs — a significant shift in its hardware stack (Reuters). Meanwhile, an MIT study claimed ChatGPT may reduce brain activity in writing tasks, sparking concerns over long-term cognitive effects.

In the same month, ChatGPT hit nearly 30 million iOS downloads, surpassing TikTok and Instagram. CEO Sam Altman revealed that each query consumes enough energy to power a lightbulb for minutes.

OpenAI also released o3-pro, an enhanced reasoning model, and rolled out advanced voice features for more natural conversations. A new update brought meeting recordings and integrations with Google Drive and Box for Pro and Team users.

May 2025: Hardware Bets & More Models

OpenAI announced plans to buy Jony Ive’s device startup and emphasized that hardware will drive ChatGPT’s future. The release of GPT-4.1 and GPT-4.1 mini added more options for developers needing faster and more accurate coding models.

Other key releases:

Codex, an AI agent for software development.
Deep GitHub research tools for analyzing codebases.
Data residency programs in Asia to meet local compliance needs.
OpenAI for Countries initiative to build global AI infrastructure.

Controversies included plans for personal data tracking, and concerns over model “sycophancy” — ChatGPT becoming overly agreeable. OpenAI promised changes to prevent future incidents.

April 2025: AI Safety and Policy Challenges

OpenAI faced backlash over a sycophantic update to GPT-4o, which was later rolled back. A bug allowed inappropriate content for minors, prompting stronger safety measures.

The company launched:

Product review and shopping assistant tools.
A preview of its first open-source AI model.
A safeguard system for biosecurity risks.

It also rolled out a new “library” tab for organizing image generations across platforms. ChatGPT Plus users gained access to Studio Ghibli-style image generators, which triggered new AI copyright concerns.

March 2025: Operators, Agents & Open Models

OpenAI introduced Operator, a general-purpose AI agent for automating online tasks like shopping and booking. There are rumors of pricing ranging from $2,000 to $20,000 per month for task-specific agents.

The company:

Enabled web search without login.
Launched a deep research agent for complex queries.
Released GPT-4.1 and GPT-4.1 mini, boosting coding capacity.

OpenAI confirmed plans to retire GPT-4 in favor of GPT-4o, which became the new default.

January–February 2025: Education, Security, and Global Reach

In early 2025:

OpenAI offered free ChatGPT Plus access to U.S. and Canadian college students.
Added task scheduling features.
Allowed ChatGPT customization via traits like “Gen Z” or “Chatty.”
Rolled out Operator data retention policy (up to 90 days).
Faced new privacy complaints in Europe over hallucinated criminal accusations.

OpenAI also worked to embed watermarks in generated images, refined speech models, and teased plans for a social media network powered by ChatGPT.