Daily Briefing · AI Models & Launches

AI Models & Launches

2:14 listen·17 stories covered
Ready to Play

AI Models & Launches — Monday, June 1, 2026

0:002:14

Full Summary

This Monday morning, NVIDIA is making headlines with multiple AI model launches, while a Chinese startup claims to outperform industry giants. NVIDIA has unveiled Cosmos 3, an open omnimodel for physical AI, confirmed by both Stock Titan and Technetbook. This model unifies vision reasoning, world generation, and action prediction, supporting text, image, video, sound, and actions. It’s designed to accelerate autonomous systems, humanoids, and self-driving vehicles, potentially reducing training time from months to days. PANews also reports NVIDIA's new RTX Spark superchip, boasting 1 petaflop of computing power for AI PCs, a massive leap from previous models. Major PC manufacturers like ASUS, Dell, HP, and Lenovo are set to launch devices with RTX Spark this fall. Additionally, NVIDIA Newsroom and the-decoder.com confirm the launch of Alpamayo 2 Super, a 32-billion-parameter model for safer, scalable Level 4 robotaxi development. Meanwhile, Chinese AI startup MiniMax has launched its M3 AI model. Both the South China Morning Post and NewsBytes report M3 can process data five times faster than its predecessor and significantly cut computational requirements, handling up to 1 million tokens of data. MiniMax claims M3 outperformed OpenAI's GPT-5.5 and Google's Gemini 3.1 Pro on a major coding benchmark. However, Tech Times notes these benchmark scores are self-reported, and the promised open weights have not yet been released. In other AI news, Fintech Singapore reports Agnes AI, a Singapore-based company, has achieved a top 10 AI lab placement on the Artificial Analysis global benchmark leaderboard with its Agnes-2.0-Flash model. Tether AI has also open-sourced TurboQuant, a memory compression algorithm that reduces the memory AI models need by up to five times, allowing devices like laptops and phones to handle larger documents and longer conversations without cloud dependency. This rapid advancement in AI models, from physical AI and robotaxis to personal AI computers and coding, means we're seeing faster, more capable AI integrated into everyday technology. Your next device could be significantly more powerful, and autonomous systems are moving closer to reality, but verifying performance claims remains crucial.

Stories Covered