News
Suno v4.5+ Released: What’s Changed
Suno v4.5+ expands music generation capabilities: now you can combine vocals, instrumentals, and the style of favorite tracks.
ChatGPT will independently choose the model for each user request
OpenAI implements automatic AI model selection in ChatGPT to improve response quality and service convenience.
Google DeepMind and OpenAI models won gold at the mathematics olympiad
Two leading language models demonstrated the ability to solve complex mathematical problems in natural language at the level of IMO…
GPT-5 is being tested for science and complex programming tasks
Evidence of internal testing of GPT-5 and the new o3 Alpha model has emerged, attracting community attention
A study showed which professions will be most affected by AI
Microsoft Research analyzed the impact of generative AI on various professions and identified who benefits the most
OpenAI urges caution when using the ChatGPT agent
Sam Altman warned about the potential dangers of the multi-step ChatGPT agent and emphasized limiting its access.
OpenAI introduced ChatGPT Agent, allowing AI to delegate complex tasks
OpenAI presented ChatGPT Agent — a tool that performs complex actions based on user requests, with additional security settings
Mistral offers the open language model Voxtral for speech
Mistral released Voxtral — an open model for recognition, transcription, and understanding of speech with multilingual support
ChatGPT to Receive Agent Mode for Working with Files and Reports
OpenAI launches Agent Mode in ChatGPT, combining office process automation and file handling in one interface
AI-generated news summaries appear in Google Discover
Google has implemented AI-generated news summaries in the Discover feed, allowing quick access to main topics.
Researchers Urge to Maintain Transparency of Thought in AI Models
Leading experts from OpenAI, Google DeepMind, Anthropic, and Meta warn of the risks of losing transparency in complex AI.
StreamDiT creates real-time video from text
StreamDiT opens up the possibility of interactive video generation and editing in real-time based on text prompts
