Current developments at ChatGPT by OpenAI (March 2025) – Image: Xpert.Digital
ChatGPT: The biggest updates at a glance
AI progress: ChatGPT is getting even better
OpenAI has made impressive progress with ChatGPT in recent months, introducing new models, enhanced features, and new pricing options. These developments demonstrate the company's commitment to making AI technology more powerful and accessible, while simultaneously creating premium offerings for demanding use cases. Here is a comprehensive summary of the key innovations.
Related to this:
New AI models
OpenAI has introduced several powerful new AI models that cover different use cases:
GPT-4.5
In February 2025, OpenAI released GPT-4.5, which was described as the “biggest and best model for general chatting.” This model offers significant improvements:
- Mia Glaese, a researcher at OpenAI, describes it as “a real step forward”
- It enables “warm, intuitive, natural, flowing conversations”
- The model demonstrates a better understanding of users' implicit expectations
- Initially available to ChatGPT Pro subscribers ($200/month), now also available to other users
Related to this:
o-Series (Reasoning Models)
In parallel with the GPT series, OpenAI has further developed its “reasoning models”:
- o3-mini is the latest small reasoning model, optimized for programming, mathematics, and scientific tasks
- It delivers faster answers than o1-mini and can match or exceed the accuracy of o1 with higher reasoning settings
- For the first time, a reasoning model is also available for free users with o3-mini
- Unlike o1, o3-mini does not support image processing
GPT-40
GPT-4o was introduced in February 2025, where the “o” stands for “omni” (Latin for “everything”):
- Twice as fast as previous models with five times better performance in text processing and speech functions
- Enhanced multimodality: Can seamlessly switch between text, speech, and images
- Improved multilingual optimization for languages such as Italian, German, and Spanish
- Improved image and video recognition and processing
- It can even sing and produce various sound effects
New user interfaces and applications
OpenAI has not only improved the models themselves, but has also developed new interfaces for interacting with ChatGPT:
Canvas
The new canvas interface offers an expanded workspace for complex tasks:
- Virtual space for writing and programming projects
- Enables targeted editing of sub-areas of a project with ChatGPT support
- It can be started manually with the command “use canvas” or opens automatically when the system detects a suitable situation
- Available for Plus and Team users, soon also for Enterprise and Edu users
Desktop application
OpenAI has released a native desktop app:
- Free ChatGPT app for macOS that allows faster access
- Offers features such as screenshot upload, voice mode, and keyboard shortcuts
- Ad-free user experience with faster access times
- Requires macOS 14 (Sonoma) and a device with Apple Silicon
- A Windows version is planned to follow
Advanced Features
ChatGPT has been enhanced with numerous useful features:
Scheduled Tasks
A new beta feature enables automated, recurring tasks:
- Users can instruct ChatGPT to perform time-based or interaction-triggered tasks
- Examples: daily news briefings, language exercises, or personal reminders
- Available for Plus, Pro and Team plans, with planned expansion to other user groups
- Users will be notified via push notification or email upon completion of the task
Extended voice communication
The “Advanced Voice” option significantly improves voice interaction:
- Support for over 50 languages in voice communication
- More natural conversations, resembling personal chats
- Custom language modes with different voice variations
- Available for “Plus” and “Team” users
Improved multimodality
The latest models can process various media types:
- Real-time recognition of images, videos, faces and emotions
- Ability to display text in 3D images
- Create video summaries
- Real-time translation between different languages
New pricing models and services
OpenAI is expanding its offerings with new premium services:
AI agents with specialization
OpenAI is planning high-priced, specialized AI agents:
- Prices range from $2,000 to $20,000 per month depending on the application
- Agent for “highly paid knowledge workers” for $2,000/month
- Developer agent earning $10,000/month
- The most expensive agent ($20,000/month) aims for “PhD-level research”
- SoftBank has reportedly committed to spending $3 billion on OpenAI's Agent products
Sora in Europe
The AI video generator Sora is now also available in Europe:
- Accessible to users in the EU, UK, Switzerland, Norway, Liechtenstein and Iceland
- Requires a ChatGPT Plus or Pro subscription
- Plus subscribers can create up to 50 prioritized videos per month
- Creates realistic and imaginative scenes from text input
Related to this:
OpenAI sets new standards with ChatGPT
The latest developments at ChatGPT demonstrate OpenAI's ongoing commitment to making AI technology more powerful, versatile, and accessible. With GPT-4.5 and GPT-4o, the boundaries of language models are being pushed further, while specialized reasoning models like o3-mini are now available free of charge for the first time.
At the same time, the company is creating better ways to interact with AI through new user interfaces like Canvas and the desktop app. The planned high-priced AI agents suggest a strategy by which OpenAI aims to serve both the mass market and premium segments.
With the expansion of language support and the availability of Sora in Europe, the technology is being made accessible to a wider international audience. These developments represent an important step in the evolution of AI assistants and demonstrate that OpenAI continues to be a leader in the development of advanced AI systems.
Related to this:
Your global marketing and business development partner
☑️ Our business language is English or German
☑️ NEW: Correspondence in your native language!
Konrad Wolfenstein
I and my team are happy to be available to you as your personal advisor.
You can contact me by filling out the contact form here wolfenstein@xpert.digital:or simply call me at +49 7348 4088 965. My email address is
I'm looking forward to our joint project.
