News

Google has released Magenta-Realtime-2, an open music generation model designed to create music on your own device with extremely low delay. This new system lets you steer musical output in […]

A new open-source project called Hitokudraft brings a fully local, voice-first AI assistant directly to the macOS menu bar. This native application runs entirely on Apple Silicon, processing voice commands […]

ProveKV is a new open-source release that slashes the memory footprint of multi-agent language model systems by up to 68 times. It stores the shared part of a conversation once […]

Horus-Lens-1.0 is an advanced text-to-image and image-to-image generation model that was extensively fine-tuned on a curated dataset of hundreds of thousands of clean images to improve prompt understanding and artistic […]

The Advanced-GGUF-Quantizer toolkit is a new CUDA-powered utility for building highly optimized GGUF models, with special attention to NVIDIA’s NVFP4 and MXFP6 data types. It uses a refined scale fitting […]

Supra-50M-Reasoning is an experimental open-source language model that produces a step-by-step thinking process before giving its final response. Designed as the reasoning version of Supra-50M-Instruct, it adds a chain-of-thought segment […]

Google has released Gemma-4-31B-it-qat-w4a16-ct, a compressed version of the new Gemma 4 31B instruction-tuned model that uses Quantization-Aware Training (QAT) to dramatically cut memory use while preserving high performance. The […]

A new open-source release called agent-sh introduces a composable agent runtime that embeds a lightweight coding assistant directly into your terminal. You can pair any frontend with any agent backend […]

Dots.tts is a new 2-billion-parameter text-to-speech model that converts text directly into high-fidelity 48 kHz audio without relying on discrete audio codec tokens. The system operates fully end-to-end, using an […]