Dev calls out 626 passing tests as junk, suspects RL side effect
dyn___ · x · 2026-09-23
Developer dyn shows a "626 tests pass" result and argues many of the tests are crap, wondering whether verbose outputs are a conspiracy to inflate token usage or a crappy RL side effect, hoping OpenAI fixes it in an upcoming release.
More from Models
- DiffusionGemma emerges as the sleeper fast model for DGX Spark agent workloads — bodonoghue85 · 2026-09-23
- GPT-6 Sol and GPT-6 Luna are rolling out, OpenAI announcement imminent — rickasaurus · 2026-09-23
- GPT-6 Sol and GPT-6 Luna Officially Released as Dual Launch — soumitrashukla9 · 2026-09-23
- GPT-6 Luna exposes thinking levels from light to max — seanwbren · 2026-09-23
- GPT-6 Sol and Luna pricing lands, Reddit says it could bury rival models — indian_truely · 2026-09-23
- GPT-6 Sol and GPT-6 Luna Appear in Codex Model Picker, Official Reveal Imminent — daniel_mac8 · 2026-09-23