Co-Founder at ARC Prize
Interviewed by MLStreetTalk to help launch ARC-AGI-2 benchmark
How media typically covers Mike Knoop
Mike Knoop as author
ARC Prize 2025 demonstrates material progress in AI reasoning systems through refinement loops, with commercial models reaching 37.6% accuracy and specialized solutions achieving 54% on ARC-AGI-2 benchmark.
“Author of "ARC Prize 2025 Results & Analysis" in ARC Prize”
Referenced in coverage
GPT-5 Pro achieves new state-of-the-art performance on the ARC-AGI reasoning benchmark, demonstrating advances in AI systems' ability to adapt to novel problem-solving tasks.
“Interviewed by MLStreetTalk to help launch ARC-AGI-2 benchmark”