NotAnotherBench
GROUNDED IN REAL HUMAN CONVERSATIONSYOUR BENCH
< R1
PROBLEMS

R1 problems

Complaints distilled from real conversations, 3,633 voices so far. Nothing here is invented.

WHAT PEOPLE COMPLAIN ABOUT

  • Sometimes makes things up (21 mentions)
  • Refuses or lectures more than people like (17 mentions)
  • Outages, errors, or limits get in the way (10 mentions)

IN THEIR OWN WORDS

  • “Verses AI (Genius) outperforms Deepseek R1 in the code-breaking challenge, being 779 times cheaper, 245 times faster, and far more accurate (Deepseek R1 solved only 45% of the challenge, while Verses AI solved it entirely).”

    SOURCE >