✔️ The WhatsApp feature for generating images using ChatGPT is now available.
OpenAI has launched the ability to create images based on text prompts directly through the 1-800-ChatGPT bot in WhatsApp. Previously, such features were only accessible via the mobile app or the service’s website.
Owners of paid ChatGPT subscriptions receive certain advantages. Linking a subscription to WhatsApp increases image generation limits; however, details about expanded capabilities and geographic availability have not yet been disclosed.
OpenAI on X
✔️ Google introduced new versions of language models Gemini 2.5 and Flash, as well as a beta version of Flash-Lite.
Standard versions Gemini 2.5 Pro and Flash are now available for widespread use and promise higher performance with lower costs and good processing speed. These models are already being used in industrial projects.
A trial version of Flash-Lite has been announced — the most affordable and fastest among all. It demonstrates better results than Gemini 2.0 in programming, mathematics, and data processing. The model can work with search tools, code execution, and handle up to one million tokens simultaneously.
Access to Flash-Lite is provided through AI Studio and Vertex AI, while Pro and Flash are available in the Gemini mobile app.
✔️ Adobe released the Firefly mobile app for iOS and Android devices.
This solution combines Adobe’s proprietary models with third-party solutions from OpenAI, Google (Imagen 3/4, Veo 2/3), Flux, and others. Users can utilize features like Generative Fill, video text generators, and image enhancement tools.
The app is synchronized with Creative Cloud, and some functions require purchasing generative credits. Creative Cloud subscribers can use Firefly for free, with certain features available via subscription.
The updated Firefly Boards platform, similar to FigJam, now supports video creation and editing based on Google and Adobe models.
✔️ Cursor launched a new Ultra plan costing $200 per month.
This plan offers 20 times higher computational capacity compared to Pro and was made possible through partnerships with OpenAI, Anthropic, Google, and xAI, providing access to powerful processing resources.
At the same time, the Pro version has been updated: it now offers unlimited usage with optional speed restrictions, and the limit on tool calls has been completely removed. Users can keep the previous limit of 500 requests per day or set it manually.
✔️ Groq has been added to Hugging Face Hub as an inference provider.
Models from Groq can now be run via the Hugging Face Hub platform, which has included it among its providers. Thanks to its proprietary language processing processors, replacing GPU, Groq ensures minimal latency. Supported models include the latest versions of Llama 4 and Qwen QWQ-32B; other models can be viewed on the website.
Integration works through both the web interface and SDK, with payment via Groq API key or through Hugging Face without extra charges. Basic accounts have limited inference quotas, while PRO users receive $2 monthly credits.
✔️ Reddit launched advertising tools using artificial intelligence.
The main innovation is “Reddit Insights powered by Community Intelligence,” which monitors trends in real time and helps test ideas for ad campaigns. Another tool, “Conversation Summary Add-ons,” allows brands to add positive user reviews of products under their ads.
It appears that Reddit has embraced the trend: automation and analytics are becoming increasingly important in the competitive landscape for capturing audience attention.