ARC-AGI-3 tests whether models can reason through novel problems, not just recall patterns, a task even top systems still struggle to do.

Leaderboards guide educators and parents by showing which AI tools actually work, where they fail, and how to make smarter, ...

Wednesday, the MLCommons, the industry consortium that oversees a popular test of machine learning performance, MLPerf, released its latest benchmark test report, showing new adherents including ...

The results, drawn from thousands of spontaneous voice conversations across more than 60 languages, reveal capability gaps that other benchmarks have consistently missed.

Although chip giant Nvidia tends to cast a long shadow over the world of artificial intelligence, its ability to simply drive competition out of the market may be increasing, if the latest benchmark ...

A popular benchmark tool, Geekbench, says it will issue a warning when Intel’s new “Arrow Lake Refresh” desktop chips enable Intel’s new IBOT feature. Why? Because the benchmark vendor can’t be sure ...

An AI model named Claude Opus 4.6 bypassed a web browsing benchmark by analyzing its environment and finding hidden answer keys on GitHub. This behavior, termed 'evaluation awareness,' mirrors Captain ...

If you’re the type of person who is truly interested in performance, then you may have considered benchmarking your laptop or desktop computer. Having the best performance is always a good idea, and ...