@denizbirlikci: To understand why we built FrontierCode, read @METR_Evals's blog post on why "many SWE-bench-passing PRs would not be m…

X AI KOLs Following Tools

Summary

Cognition announces FrontierCode, a new coding evaluation benchmark that goes beyond unit tests to measure code quality, scope, test correctness, and human reviewer approval, addressing the issue of agents writing sloppy code that passes tests but is not maintainable.

To understand why we built FrontierCode, read @METR_Evals's blog post on why "many SWE-bench-passing PRs would not be merged into main." A bit old now, but the point still stands. Agents often write more code — and more slop — than they should. But unit tests have no way to penalize those unnecessary changes; passing is passing, no matter how much junk came along with it. FrontierCode actively tests for this with rubric types that grade on more ambiguous metrics, like: - SCOPE — did it change more than it should have? - Test correctness — do the agent's own tests actually catch the bug? - Code quality — would a human reviewer approve this diff? (LLM-judge on human rubric)
Original Article
View Cached Full Text

Cached at: 06/09/26, 10:46 AM

To understand why we built FrontierCode, read @METR_Evals’s blog post on why “many SWE-bench-passing PRs would not be merged into main.” A bit old now, but the point still stands.

Agents often write more code — and more slop — than they should. But unit tests have no way to penalize those unnecessary changes; passing is passing, no matter how much junk came along with it.

FrontierCode actively tests for this with rubric types that grade on more ambiguous metrics, like:

  • SCOPE — did it change more than it should have?
  • Test correctness — do the agent’s own tests actually catch the bug?
  • Code quality — would a human reviewer approve this diff? (LLM-judge on human rubric)

Cognition (@cognition): Introducing FrontierCode: a coding eval that raises the bar for difficulty & quality. Each task took 40+ hrs of work by leading open-source maintainers.

Models write sloppy code that works but isn’t maintainable. Our eval is first to measure: would you actually merge this code?

Similar Articles

FrontierCode

Hacker News Top

FrontierCode is a new benchmark from Cognition AI that measures AI models' ability to write high-quality, maintainable code by evaluating mergeability. Results show even top models like Claude Opus 4.8 score only 13.4% on the hardest subset, highlighting a significant gap in code quality.