LeadAIExplainer
What "671 billion parameters" really means: mixture-of-experts explained
Big AI models now quote two sizes, total and active. Here is how mixture-of-experts works and why the gap between the two numbers matters.
AI · Explainer
How to read an AI benchmark: ECI, Arena and the numbers in launch posts
Launch posts quote test scores, arena ranks and composite indexes. Here is what each one measures, how it goes wrong and what to ask.
AI · AnalysisBest AI for coding in 2026: what the benchmarks show and what they miss
The top coding models are within a few points of each other, and the scores come mostly from the companies selling them. Here is how to read them.