Open large language model sets new multilingual benchmark
A new open LLM matches state-of-the-art performance across 30+ languages while remaining lightweight enough for edge deployment.
Curated updates from AI research labs, companies, and regulators. Use the controls below to explore by category or focus on top stories.
A new open LLM matches state-of-the-art performance across 30+ languages while remaining lightweight enough for edge deployment.
Regulators from multiple regions publish a shared set of AI safety principles covering evaluation, transparency, and incident reporting.
A robotics team demonstrates a unified model that can pick up unseen objects with minimal additional training.
A new survey finds that over 60% of organizations piloting AI copilots report measurable productivity improvements.
A new research paper introduces benchmarks for AI agents that act over extended periods and across multiple tools.
Practical bias mitigation methods are tested on large-scale recommendation and ranking systems with promising results.