Open models still can't handle CTF-style CVE hunting, closed models lead
xeophon · x · 2026-10-22
In a discussion on open-model cyber capabilities, xeophon argues that tasks requiring many chained reasoning layers—finding CVEs and php2shell-style exploits—are currently within reach of closed models only. He says current open models are not good enough for these CTF-style tasks, a view he validated after talking with the researcher involved, noting cyber benchmarks also test this.
More from Models
- Axion Labs launches non-transformer decision models for efficient edge reasoning — tw_killian · 2026-10-23
- Mistral Research Head on Le Chonk: Trained on 3,800 GPUs, 10-Trillion-Parameter Models Coming — mattturck · 2026-10-23
- Northwestern paper decomposes LLM decision errors into belief loss vs. optimization residual — JessicaHullman · 2026-10-23
- Researcher flags Claude's tiring 'animism' speech quirk: chapters ask, benchmarks question — jonasgeiping · 2026-10-23
- Fastino's GLiDE v1.2 tops HF Decision Index at 71.5, adds multimodal inputs — multimodalart · 2026-10-23
- Claude Can't Stop Writing Sycophantic Comments, Devs Say Gemini Astra Is Cleaner — ctjlewis · 2026-10-23