AI Intel

New research explores how to determine if a multi-agent code judge is "grounded" using two label-free measurements, proposing a judge that refrains f…

HybridInfer proposes a thermal-aware reinforcement learning approach for routing LLM inference across on-device, edge, and cloud tiers to optimize pe…

Parley is introduced as a federated, decentralized chat system that offers compatibility with the plain IRC protocol.

OpenAI's GPT-3 model has officially been discontinued, marking the end of an era for many early adopters of modern language models.

A Financial Times report indicates that corporate America is increasingly rejecting expensive frontier AI models in favor of more cost-effective open…

New research proposes an architectural mediation approach, using the Model Context Protocol, to bridge LLM agents and data spaces for sovereign and g…

SaleSmartly, an AI-powered tool designed to convert conversations across various channels into customers, has launched on Product Hunt.

The latest llama.cpp server update introduces RANK pooling batch splitting to improve efficiency for causal LLM rerankers like Qwen3 and Qwen3-VL.

Claude Code added x-claude-code-prompt-id to gateway hint headers, allowing LLM gateways to group requests related to a single user prompt for better…

Qwen's company has quietly released "decision-model-preview," a potential competitor to Jev, although only a documentation page is currently availabl…