# Let a cheap model fix real bugs overnight, and most of its PRs didn't survive review — retained record

> Retained reddit record from reddit:AI_Agents.

Canonical: https://highsignal.app/data/records/fb0d2492005e26b1

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:AI_Agents)  published 2026-10-03 09:15 ingested 2026-10-04 02:35

[original source ↗](https://www.reddit.com/r/AI_Agents/comments/1wwisl2/let_a_cheap_model_fix_real_bugs_overnight_and/)

## retained text



```
This is my own setup, my own repos, and I'm a bit stuck on what to do with the result.

An audit turned up some serious bugs. Instead of fixing them myself I let a cheap, fast model (DeepSeek, during an API discount window) grind through them overnight.

Round 1 was 14 PRs. I had a stronger model (GPT-6.1 Sol) re-review every one of them. 6 got merged after all CI passed. I closed the other 8.

The reasons were things like a fix that broke other valid inputs, a test that needed a tool our CI doesn't have, a bypass that still worked, a key test that never even ran in CI, and a new test that couldn't catch the bug even when I deliberately broke the code.

So for round 2 I flipped it. A stronger model designs each fix, another one reviews the design, and the cheap model only implements. That's 7 PRs so far and none approved yet.

Where do cheap-model fixes fail for you? And what split between planning, implementing and reviewing has actually worked?
```



document fetched 2026-10-04 00:19 · canonical https://reddit.com/r/AI_Agents/comments/1wwisl2/let_a_cheap_model_fix_real_bugs_overnight_and
