Models Are Getting Dumber on Purpose

Anthropic: Introducing The Conceptual Reasoning Index

DeepSeek V4 Flash 0731

When AI Benchmarks Plateau: A Systematic Study of Benchmark Saturation

What's the largest software project AI can complete on its own?

Any text-to-SQL benchmark should address difficulties of real-world data stores

Are AI Labs Pelicanmaxxing?

Kimi K3, and what we can still learn from the pelican benchmark

The gap between open weights LLMs and closed source LLMs

GLM-5.2 is the new leading open weights model on Artificial Analysis

FrontierCode

Claude Opus 4.6 accuracy on BridgeBench hallucination test drops from 83% to 68%