By skyforbes 
windows and began transforming into full-fledged platforms. OpenAI introduced applications directly within ChatGPT, Spotify opened the door to personalized recommendations, Google allegedly prepares its response in the form of Gemini 3, and Elon Musk demonstrated that his Grok can generate videos in under 15 seconds.
Applications in ChatGPT
OpenAI launched a revolution in October 2025 in the form of applications running directly inside ChatGPT conversations. As of October 6, users can invoke interactive applications from companies like Spotify, Canva, Figma, Zillow, or Coursera during chat without having to leave the conversation window. CEO Sam Altman presented it as “a new generation of apps that are interactive, adaptive, and personalized” that should turn ChatGPT into a true productivity platform.
It works surprisingly intuitively — just mention the app name in conversation, and ChatGPT automatically invokes the corresponding mini-program that displays directly in the chat. In demos, OpenAI employees showed how during a single conversation they asked Canva to create a dog-walking poster, then generated a pitch deck, and then searched through Zillow for available homes in Pittsburgh with an interactive map right inside the conversation. For developers, a new Apps SDK is available that allows creating custom applications on open standards.
This strategy represents OpenAI’s attempt to build an ecosystem around ChatGPT that would function similarly to an operating system. Unlike the previous GPT Store, where users had to actively visit a separate app store, the new solution pulls applications directly into the normal conversation flow, which ensures better distribution for developers and a more natural experience for users. In the coming weeks, additional services like DoorDash, Opentable, Target, and Uber will join the offering.
Spotify Personalization in ChatGPT
Spotify became one of the first major platforms on October 6, 2025, to bring its integration into ChatGPT as a full-fledged application with access to personalized recommendations. While regular users (Free tier) can browse existing playlists like Discover Weekly or New Music Friday, Premium subscribers gain the ability to have completely new, personalized song selections generated directly tailored to their prompts. Just mention Spotify in a ChatGPT conversation, connect your account, and ask for a playlist for a weekend road trip or the latest hits from a favorite K-Pop star.
The integration works in 145 countries, in English, and is available to all logged-in ChatGPT users (Free, Plus, and Pro) on the web and mobile apps for iOS and Android. Spotify brings its legendary personalization technology to the system, which it combines with the expertise of its human editors. The result is an experience that can respond to mood, genre, or specific theme — whether searching for podcasts by guest or Latin artists on repeat.
Regarding privacy, Spotify explicitly assures that it does not share music, podcasts, or other audio or video content with OpenAI for model training purposes. Users have everything under control — connections can be disconnected at any time, the entire system is opt-in. For artists and podcasters, this means a new platform for discovering their work, which could change the way people search for new music — instead of clicking in the app, they simply describe what they want to hear.
Google Gemini 3
Google allegedly prepares its biggest AI bomb of the year in the form of Gemini 3 model, which could arrive as early as the second half of October 2025. According to information from the developer community and A/B testing on AI Studio, the most likely date appears to be October 9, which would align with Google’s historical pattern of releasing major models at year-end. If the date is confirmed, Gemini 3 would arrive just days after OpenAI DevDay and the introduction of apps in ChatGPT — Google would thus catch a response to GPT-5, which OpenAI launched in August 2025.
Unofficial benchmarks and code leaks suggest that Gemini 3 Pro could significantly surpass GPT-5 in advanced reasoning — according to tests on the so-called “Humanity’s Last Exam,” Gemini 3 achieved a score of 32.4% compared to 26.5% for GPT-5. The model is expected to bring advanced multimodality including real-time video analysis up to 60 FPS, expanded context window to millions of tokens (equivalent to entire books in a single prompt), and integrated “Deep Think” mode for complex problem-solving. References to “gemini-beta-3.0-pro” and “gemini-beta-3.0-flash” have already been discovered in Google CLI repositories.
The planned rollout should occur in phases — first limited access for enterprise partners and Vertex AI (October to December 2025), then public API for developers in early 2026, and subsequently complete consumer launch probably in Q1 2026, possibly together with new Pixel devices. If Google actually meets the October deadline, it will be one of the most significant milestones in the 2025 AI race.
Grok Imagine v0.9
Elon Musk unveiled Grok Imagine v0.9 on October 6, 2025, a new version of his AI generator that aims directly at OpenAI and their Sora 2. The main trump card is speed — according to Musk’s words, Grok Imagine can generate video in under 15 seconds, making it one of the fastest AI video generators on the market. The update comes just days after the launch of Sora 2, showing how intense the competition is currently in the AI content generation area.
Version 0.9 brings instant generation of text, images, and video within a single system. Musk emphasized on platform X that it’s not just about video — Grok 4 Fast ensures the fastest text generation and the entire system now supports a voice interface that should make interaction even easier for users. It’s a strategic move by which xAI (Musk’s AI company) tries to compete not only with OpenAI but also with other players like Google or Anthropic.
The main innovation is the voice-first interface, which allows users to open the application and immediately start speaking without any typing. Just activate “Open App in Voice Mode” in settings and Grok is ready to respond immediately. This approach corresponds to a broader trend in AI assistance — movement toward more natural, multimodal interaction, where it’s not just about text but a complete combination of voice, text, images, and video in one tool.
(see the generated image above)
Sources:
- https://currently.att.yahoo.com/att/openai-just-turned-chatgpt-app-204723722.html
- https://newsroom.spotify.com/2025-10-06/spotify-personalized-prompts-chatgpt/
- https://www.digitaltrends.com/computing/googles-answer-to-gpt-5-might-be-days-away-as-gemini-3-launch-touted/
- https://www.newsbytesapp.com/news/science/xai-upgrades-grok-imagine-for-faster-more-realistic-ai-videos/story