Benchmark: AI doesn't find bugs unless you tell it what's wrong

Details

External ID

49923102

Source

HN

Company

—

Product

Benchmark: AI doesn't find bugs unless you tell it what's wrong

Website domain

swesweep.com

Launched

Oct. 1, 2026

Cohort

—

Upvotes

5

Upvotes percentile

0.11764705882352941

Tags

—

Fetched at

Oct. 2, 2026, 1:01 a.m.

Updated at

Oct. 2, 2026, 1:01 a.m.

Description

New benchmark from researchers at Meta, Stanford, Harvard, UW, including the researchers who've worked on SWE-bench, ProgramBench etc.Most benchmarks just test if AI can fix a problem you've already pointed out.But obviously it would be much better to fix problems before you or any user runs into it. Like, isn't it crazy that we still have to wait for people to open tickets before a lot of obvious bugs get found?We wanted to test that capability at scale. Turns out that models still are terrible at it (best setup we tested still fixed < 5% of bugs)We have 100 repos of 22 languages and 4k bugs between. All the bugs are real-world bugs from github. We do a lot of filtering to ensure everything can be solved in this setting.===========================================Sol 5.6 (xhigh) 4.7% $7,230Luna 5.6 (xhigh) 2.5% $224Terra 5.6 (xhigh) 1.5% $357Luna 5.6 (high) 1.4% $28Opus 5 (xhigh) 1.3% $5,363Kimi K3 0.6% $2,451Luna 5.6 0.5% $4GPT-5.4 Mini (high) 0.5% $122GPT-5.4 Mini 0.2% $5Gemini 3.5 Flash Lite 0.1% $6===========================================Also the best model is very expensive.We have a lot more FAQ on the website https://swesweep.com/ Oh and we're all open-source (MIT license) at https://github.com/facebookresearch/swe-sweepCurious what you all think!

• Source listing
• Website