Google has launched an experimental feature that enables users to leverage the latest Gemini models to generate audio overviews for appropriate search queries. The overviews can be played within the search results page on a player that displays selected relevant web pages.
Researchers have developed a new digital restoration method that leverages deep learning techniques to reconstruct damaged areas and uses printable polymer films to physically apply the digital restorations onto the damaged paintings without altering the original canvas and pigments.
Spain-based Multiverse Computing raised €189 million Series B led by Bullhound Capital to scale its CompactifAI technology that compresses large language models by up to 95% while maintaining performance, enabling AI to run efficiently on edge devices and reducing inference costs by 50-80%.
Anysphere, the maker of the popular AI code editor Cursor, has announced a $900 million Series C at a $9.9 billion valuation. Thrive, Accel, Andressen Horowitz, and DST participated in the startup's funding round.
Paris-based Moments Lab raised a $24 million Series B to expand its AI-powered video production tools that help Hollywood studios and media companies create content up to seven times faster by automatically indexing footage and generating rough cuts from text prompts.
Edinburgh-based Wordsmith AI raised $25 million Series A led by Index Ventures to scale its AI agent platform that transforms legal departments from bottlenecks into "revenue accelerators" by embedding legal intelligence across business workflows.
Microsoft has launched Bing Video Creator, a free AI-powered tool that uses OpenAI's Sora technology to generate 5-second videos from text prompts, now available on mobile, and with desktop support coming soon.
DeepSeek has released an updated version of its R1 reasoning model and a new distilled 8B variant (DeepSeek-R1-0528-Qwen3-8B) that achieves state-of-the-art performance on the AIME 2024 competition math benchmark.
Black Forest Labs has launched FLUX.1 Kontext, a suite of AI image generation models combining text-to-image creation with real-time image editing through text instructions. The model supports iterative editing that preserves character consistency at speeds up to 8x faster than competitors.
Perplexity has launched Labs, a new AI tool for Pro subscribers that creates comprehensive work deliverables like reports, dashboards, and web apps through sustained automated research, coding, and content generation.
Anthropic has launched Claude 4 with two new models: Opus 4, which it claims is the world's best model for coding, and Sonnet 4, which builds on Sonnet 3.7's already impressive capabilities. Additionally, the company announced its coding agent, Claude Code, is now generally available.