Redis Creator antirez Slams Coding Benchmarks as Mostly Garbage
Redis creator antirez criticized top coding benchmarks as largely misaligned with real-world model performance, arguing most test questions are low quality and mainly probe tool-calling friction. He suggested major labs rely on private internal benchmarks to truly measure progress.
2026-09-03 ~ 2026-09-03 · 2 related posts
- antirez rips top coding benchmarks: mostly 'crap' testing tool friction, not real ability — antirez · 2026-09-03
- antirez: coding benchmarks are 'mostly crap' and badly misaligned with how models train — antirez · 2026-09-03