Microsoft’s MAI-DxO AI achieved 85.5% accuracy in diagnosing complex cases—far outpacing doctors and lowering diagnostic costs in trials.

Google's Gemini CLI brings powerful, free AI assistance to developers' terminals with open-source access and seamless code editor integration.

Tencent's new open-source Hunyuan3D-2.1 delivers photorealistic 3D models, setting a new standard in accessible, AI-powered design.

Hailuo AI, by MiniMax, offers free, registration-free AI-powered video generation—redefining how creators, marketers, and educators work.

MiniMax AI introduces MiniMax-M1, a 456B parameter open-source LLM with a 1M-token context, setting new benchmarks in large language models.

Meta’s V-JEPA 2 enables robots to learn object recognition and planning from scratch using self-supervised video training.

Google’s Veo 3 creates ultra-realistic, audio-synced videos from text. The new Flow tool also brings cinematic AI filmmaking to a wider audience.

Google’s Gemma 3n, unveiled at I/O 2025, brings cloud-free, multimodal AI to low-resource mobile devices, powering advanced on-device intelligence.