Chinese artificial intelligence developer DeepSeek today released a new series of open-source large language models. V4, as ...

April 30, 2026 expert reaction to study evaluating performance of a large language model on the reasoning tasks of a physician . A study published in Science evaluates the perform ...

DeepSeek says both models are more efficient and performant than DeepSeek V3.2 due to architectural improvements, and have ...

A wave of 2026 developments — from Anthropic's Model Context Protocol to Microsoft's GraphRAG concept and rigorous benchmarks like Terminal-Bench 2.0 and SWE-Bench Pro — is redefining how AI teams ...

A Harvard-led study found that OpenAI’s o1-preview large language model matched or exceeded hundreds of physicians in six clinical reasoning experiments, particularly excelling in emergency triage.

A new benchmark released by Simbian is challenging one of the most widely held assumptions in artificial intelligence: that the same models capable of finding vulnerabilities can also defend against ...

Differential diagnosis was less accurate than diagnostic testing, but final diagnosis and management were more accurate.

As Big Tech pours unprecedented resources into scaling large language models, critics argue that transformer-based systems ...

ShengShu Technology today announces Motubrain, a World Action Model that replaces multiple task-specific systems with a single, unified model that functions as a robotic brain for the physical world.

Built on more than 180m real patient interactions, validated by U.S.-licensed clinicians and now benchmarked against every leading frontier model, Polaris 5.0 leads safety, compliance and empathy for ...

This isn't about rejecting large models; it's about having the engineering discipline to use smaller, specialized models ...