Cisco’s open Antares SLMs tackle vulnerability localization with 350M to 3B parameters
huggingface · x · 2026-07-24
Cisco’s Antares is a family of open small language models for vulnerability localization, with sizes of 350M, 1B, and 3B parameters.
- The post argues that small open models can detect security vulnerabilities, not just large closed models.
- Antares works by iteratively narrowing down a repository: it starts from the vulnerability description, searches for relevant patterns, reads candidate files, incorporates new evidence, and changes direction when needed.
- The accompanying chart positions Antares against other models on file-level F1, showing strong performance for its size.
Related event: Cisco Open-Sources Antares Small Models for Code Vulnerability Localization(4 posts)→
More from Research
- Open-source decision models sold their own base in Red Alert 2 — then learned until no human could beat them — heathercmiller · 2026-10-27
- DeepMind paper: LLMs can't sample randomly on their own — pass random numbers from code instead — rohanpaul_ai · 2026-10-27
- World Model Assigns Weight and Friction to Objects, Then Simulates with a Physics Engine — jonstephens85 · 2026-10-27
- Token Collapse: Study Finds Graph Soft Prompting Tokens in LLMs Fail to Encode Structure — Sapienza · 2026-10-27
- Mathathon 2.0 launches with Terence Tao's SAIR, $1.5M in compute for teams — _sathvikr · 2026-10-27
- The 'sweet lesson' is real: two kinds of ML problems need opposite strategies — mike64_t · 2026-10-27