AI Intel

Microsoft's open-source tools were compromised in an attack that led to the theft of AI developers' passwords.

Xiaomi announced achieving over 1,000 tokens per second output on its 1 trillion parameter MoE model, MiMo-V2.5-Pro UltraSpeed, using a single standa…

Krisp has launched a new Voice Translation API that provides real-time speech-to-speech translation capabilities.

ZeroGPU has been introduced as a compute-efficient layer designed to optimize AI inference processes.

New research explores the evaluation of hallucinations in domain-adapted Large Language Models (LLMs), specifically focusing on fine-tuning Llama-2 w…

Google has made its image generation models, Nano Banana 2 and Nano Banana Pro, generally available to developers and businesses through GoogleAI Stu…

Grok Build is now available in Beta for SuperGrok and X Premium+ users, offering Plan Mode, image/video creation with Imagine, and automation buildin…

Mistral Vibe has been officially shipped, becoming an AI agent designed for long-horizon productivity and coding, featuring Work mode, Code mode, a C…

Mistral AI has launched its Connectors API in Public Preview, allowing developers to register MCP connectors once and reuse them across Le Chat, AI S…

Google has released Gemma 4 12B, a unified, encoder-free multimodal model under an Apache 2.0 license, designed for high-performance intelligence dir…