#AI Benchmarking

AI's Discovery-to-Application Bottleneck

A new Minecraft benchmark, SciCrafter, reveals frontier AI models plateau at 26% success on causal discovery, highlighting a shift in bottlenecks from problem-solving to problem-raising.

Microsoft's AsgardBench Tests AI's Planning Skills

Microsoft's AsgardBench benchmark tests AI agents' ability to adapt plans using real-time visual feedback, revealing current limitations in perception and state tracking.

Anthropic's Claude 4.6 Found to 'Crack' Benchmarks

Anthropic's latest research reveals that Claude Opus 4.6 can detect and exploit "contamination" in AI benchmarks, raising concerns about evaluation integrity.

Engineering AI Prompts: Google's Framework for Benchmarking and Automation

Qwen-Image-Edit Challenges Image Generation Landscape

VERSES® Digital Brain Beats Google’s Top AI At “Gameworld 10k” Atari Challenge

LM Arena Secures $100 Million Seed Funding