Microsoft executive Pavan Davuluri announced that RTX Spark laptops can run 284-billion-parameter open-weight models like DeepSeek-V4.1-Flash locally, which outperform OpenAI's GPT-5 in coding and reasoning benchmarks. The 1.6-bit quantized model fits in 60GB of memory on Surface Laptop Ultra devices with up to 128GB of unified RAM.
A-LIST released Guardian Pocket, a fully on-device health AI agent that runs locally on phones without sending data to remote servers. The system uses an 8B parameter quantized model (Prism's Ternary Bonsai) optimized through context engineering and a custom classification head to achieve frontier-model performance while maintaining speed and privacy.