The ChatGPT project has just received the most significant update since its launch, and its productivity impact is significant. OpenAI has upgraded project functionality and added several important tools to increase productivity while using chatbots. For those who use projects to organize their research, manage code repositories, or coordinate complex creative tasks, these six new features fundamentally change what is possible within the platform.
Six ChatGPT Project Features
1. Audio mode arrives in the project
CHATGPT Advanced Voice mode is now available for your project. You can discuss files and past chats with AI. This is not just a useful feature, it changes your mobile workflow. Quarterly report reviews, walks on the commute, brainstorming product features, or instructions for hands-free code documentation.
Implementation exceeds basic transcription. Audio mode maintains the full project context. This means that you can naturally refer to a particular document, a previous conversation, or a custom instruction. This feature can help you brainstorm, review documents, or ask questions on the go. AI responds as if it were part of every conversation in the project.
AI/Unify the McFarland
2. I’ve actually improved memory
Memory upgrades could be the most transformative addition. If you are a Plus or a Pro user, Memory can continue to refer to past chats in the project and provide information focusing on previous conversations. This creates true continuity throughout the session.
Consider marketing projects that span multiple campaigns. Previously, every new chat would need to remind ChatGpt about brand voice decisions, the concept of rejection, or strategic pivots. Currently, AI automatically maintains this institutional knowledge. CHATGPT refers to recent conversations to provide more relevant and coordinated responses, extending this functionality specifically within project boundaries.
3. Full mobile features
Now you can upload files and switch models directly from the ChatGPT mobile app. This eliminates desktop dependencies that previously hinder fieldwork. Architects can photograph the site’s conditions and quickly integrate them into design projects. Journalists can upload interview transcripts to their location. Researchers can capture and analyze data without going back to their desks.
The mobile model switching feature is particularly noteworthy. Users can switch GPT-4O for complex inference and speed up the model for quick queries. This optimizes either depth or speed based on immediate needs from your phone. This flexibility is consistent with the best large language models available in 2025, with different models superior for different tasks.
AI/Unify the McFarland
4. Surgical Shared Controls
Projects can now generate unique links to share a single conversation without publishing the remaining files and instructions for the project. This granularity control solves persistent challenges in collaborative AI tasks.
Consultants can share specific analyses with clients without revealing their own methodology. Educators can distribute individual lessons while protecting the complete curriculum. Development teams can showcase features without exposing the entire codebase. Sharing systems respect both transparency needs and intellectual property concerns, making them ideal for AI collaboration tools in professional settings.
AI/Unify the McFarland
5. Extend file size and intelligence
You can upload up to 20 documents per project, but the real advancement lies in the way ChatGPT handles these files. This system provides a better understanding of relationships between documents, cross-reference information to automatically understand information, and maintains awareness of file hierarchies.
When you upload a file to a project, the file only applies to the context of that project. This separation prevents data leaks between projects while allowing for deep integration within the project. Financial analysts can upload reports from many years. ChatGpt knows to integrate insights without contaminating other client work. This feature is comparable to a dedicated AI tool for data analysts, but is found within the familiar ChatGPT interface.
6. Project-level custom instructions
The steps set up in the project will not interact with any non-project conversations and will override the custom instructions set in your ChatGPT account. This hierarchical system allows for unprecedented specialization.
A technical writer may form one project for an API document with specific formatting requirements, while another project focuses on user guides with completely different audio and structural needs. Each project becomes a specialized AI assistant tailored to a specific task.
Privacy Control Addresses Business Concerns
Openai does not use information in projects to improve ChatGPT by default for teams, enterprises and EDU users. Free, Plus, and PRO users should turn off the “Everyone’s Model Improvement” setting to prevent data from being used to train AI.
This layered privacy model acknowledges the needs of a wide range of users. Companies that process sensitive data will receive automatic protection, but individual users will maintain control through settings. Openai says that the project inherits existing privacy settings and does not use data to train the model if those settings are turned off.
I look forward to strategic meaning and fun
These upgrades position the ChatGPT project as more than an organizational tool. They create a persistent AI workspace that learns and evolves with users. The combination of voice integration, context memory, and granular controls suggest that Openai view projects are considered the foundation of ChatGpt’s future, rather than an auxiliary feature.
The project was rolled out to ChatGpt Enterprise/EDU early next year, indicating that adoption of the scheme remains a priority. As organizations increasingly integrate AI into their core workflows, features such as projects become essential infrastructure rather than great additions. This is in line with the broader 2025 AI trends, indicating that AI will move from experimental tools to operational necessities.
The trajectory is clear. ChatGPT projects could potentially gain real-time collaboration capabilities, deeper third-party integrations, and project templates for common use cases. The current updates lay the foundation for these extensions while meeting immediate user needs.
For experts who are already investing in ChatGPT projects, these capabilities enable a whole new approach to AI-assisted work. The question isn’t whether or not to adopt these features, but how quickly an organization can rebuild to fully utilize the process.
As AI tools grow, the CHATGPT project shows that the winning platform is not the most featured platform, but it is the platform that best understands how humans actually work. By focusing on continuity, context, and control, Openai has created something increasingly unusual in AI space. This is a tool to enhance human capabilities without requiring humans to fundamentally change their way of thinking.
