# OpenAI finds roughly 30 percent of popular AI coding test is broken

> OpenAI finds roughly 30 percent of popular AI coding test is broken Key Points - OpenAI is pulling its endorsement of the AI coding test SWE-Bench Pro after a review found roughly 30 percent of its tasks are flawed.

- **Source:** [The-decoder](https://the-decoder.com/openai-finds-roughly-30-percent-of-popular-ai-coding-test-is-broken/?utm_source=gearopen.com&utm_medium=referral&utm_campaign=feed&utm_content=6a4fbe8672fc80d7c7913a77)
- **Published:** 2026-07-09
- **Category:** AI & Bots
- **Tags:** #openai #ai
- **Canonical:** http://localhost:4321/a/6a4fbe8672fc80d7c7913a77

---

_GearOpen's distilled summary. The full original article is published at The-decoder (source link above); GearOpen does not republish the source's full text._