Topic
Benchmarks
3 stories about Benchmarks that passed the news check, newest first.
Research ·
The most misunderstood graph in AI just got a major update
An AI evaluation nonprofit, METR, published a report finding that some AI capabilities continue to improve along an exponential trajectory.
Artificial intelligence – MIT Technology Review
Release ·
Introducing GPT-6.2 Astra and Sol
OpenAI introduced GPT-6.2 Astra, described as its new frontier model for the hardest projects, and GPT-6.2 Sol, described as an efficient, versatile model for everyday work.
OpenAI News · 2 sources
Research ·
Diagnostic probes reveal shortcuts induced by the benchmark protocol in vision-language model annotation pipelines for video-based self-harm detection
A study reports diagnostic probes revealing shortcuts induced by the benchmark protocol in vision-language model annotation pipelines for video-based self-harm detection.
Medical research : nature.com subject feeds