Hacker News15h ago571 points267 commentslive
GLM 5.2 beats Claude in our benchmarks
Semgrep's IDOR benchmark shows GLM 5.2 (39% F1) beats Claude Code (32%) and Claude Opus 4.8, highlighting that open-weight models are now competitive for security tasks without heavy harnessing.
Read the full story atsemgrep.devWhy this is in the Signal
LAXIMA AI Signal curates the highest-velocity stories across Hacker News, GitHub trending, and new Hugging Face / Replicate model releases — quality-filtered, deduplicated, and refreshed every four hours. This item surfaced from Hacker News with 571 points (by jms703). We link straight to the original source above — see the full live feed.
More AI Signal briefs
- HNAI boom risks global financial crash, warn central bankers
- GHAgent behavior clone for browser using, targeting general GUI using and distributed trajectory collecting.
- HNFord rehires 'gray beard' engineers after AI falls short
- HNI used Claude Code to get a second opinion on my MRI
- HN5k Restaurant Menus, Years 1880-1920
- HNEU Open Sources Ten-Year Network Development Planning Tools
- HNFlock cameras track more than your license plate, and they're spreading fast
- HNEU to legislate about Chat Control behind closed doors