OpenAI upgrades ChatGPT with GPT-5.6 Luna, cutting factual errors by 62% and slashing prices at an unprecedented pace as competition from cheaper Chinese AI models intensifies.
Xiaomi open-sources its robotics AI foundation model, Xiaomi-Robotics-1, trained on over 100,000 hours of data, offering a rare end-to-end blueprint in a proprietary-dominated field, but missing key details like model architecture and parameter count raise serious questions about its real-world utility.
Anthropic is building a custom AI chip design team to co-design hardware and models, as surging demand for Claude outpaces existing infrastructure partnerships with AWS, Google, Nvidia, and AMD, with Samsung being scouted as a potential manufacturing partner.
DeepSeek warns developers that significant API price hikes are coming, though the company has yet to reveal new rates or an effective date, leaving users on notice that costs will soon rise substantially.
Google DeepMind shakes up its leadership structure, elevating Koray Kavukcuoglu to SVP and shifting Demis Hassabis to chairman, as the tech giant scrambles to close a widening gap with OpenAI and Anthropic in the AI race.
Meta AI launches Muse Code in beta, a terminal-based coding agent powered by the new Muse Spark 1.2 model that can autonomously plan, write, and validate code across large repositories, with benchmark tests showing it successfully optimizes GPU kernels over 1,000+ tool calls for significant performance gains on NVIDIA Hopper …
Google shakes up its AI leadership as Demis Hassabis steps back from day-to-day operations at DeepMind to focus on AGI strategy, Koray Kavukcuoglu takes the reins overseeing Gemini's 950 million users, and legendary engineer Jeff Dean departs after 27 years to co-found a new AI-focused public benefit corporation with Google …
NVIDIA launches Alpamayo 2 Super, a commercially available open reasoning model for autonomous vehicles that ranks first on all major autonomous driving benchmarks, outperforming Gemini 2.5 Pro and GPT-4o, while offering developers full fine-tuning and deployment rights under a permissive license.
Liquid Foundation Models (LFMs) launch as a powerful new class of multimodal AI supporting text, vision, and audio, optimized for fast on-device and edge deployment across major inference runtimes with flexible formats including GGUF, MLX, and ONNX.
Mistral launches Shieldstral, a compact 3-billion-parameter AI safety model that accepts natural language policies at inference time, outperforms models seven times its size including OpenAI's GPT-OSS-Safeguard, runs on a 16GB GPU, and is released open-weight under Apache 2.0.