Mystery Kaggle leaderboard team revealed, speculated to be OpenAI or Anthropic testing agents

JFPuget · x · 2026-09-04

The code-named team leading a Kaggle competition was revealed to be @aimalysheva's crew. The community had speculated it was OpenAI or Anthropic secretly evaluating their agents on live competitions — possibly deleting accounts when agents underperformed — highlighting Kaggle as a real-world agent benchmark.

Original post →

More from AGI Musings

AGI Musings channel →