Antares: Foundation Models for Agentic Vulnerability Localization
Supriti Vijay, Aman Priyanshu, Didier Chapoteau, Arthur Goldblatt, Jianliang He, Kimia Majd, Fraser Burch, Baturay Saglam, Takahiro Matsumoto, Zhuoran Yang, Amin Karbasi
cs.CR, cs.AI
Submitted: 2026-08-03
Comments: 57 pages, 12 figures, technical report for antares
Code: https://github.com/cisco-foundation-ai/vulnerability-localization-benchmark
License: http://creativecommons.org/licenses/by/4.0/
The gist: Vulnerability localization is a fundamental step in software security, requiring models to reason over large codebases and iteratively identify vulnerable implementations.
Terminology
Abstract
Vulnerability localization is a fundamental step in software security, requiring models to reason over large codebases and iteratively identify vulnerable implementations. We present Antares, a family of compact language models (350M, 1B, and 3B parameters) for agentic vulnerability localization. Based on IBM Granite base models, Antares is trained through a two-stage pipeline that combines supervised fine-tuning on cybersecurity reasoning and repository exploration data with reinforcement learning from verifiable rewards over vulnerable repositories. Across extensive evaluations, Antares-3B approaches GPT-5.5 while outperforming open-weight models over 200x larger in size. The Antares family further enables fast, low-cost local inference, completing a full 500-task evaluation sweep in approximately 15 minutes on a single H100 GPU, corresponding to an amortized evaluation time of under 2 seconds and less than 0.002 per task.
Sources
- CTIBench: A Benchmark for Evaluating LLMs in Cyber Threat Intelligence
- AgenticSCR: An Autonomous Agentic Secure Code Review for Immature Vulnerabilities Detection
- RaDeR: Reasoning-aware Dense Retrieval Models
- SWE-Bench Pro: Can AI Agents Solve Long-Horizon Software Engineering Tasks?
- RepoAudit: An Autonomous LLM-Agent for Repository-Level Code Auditing
- Matching Features, Not Tokens: Energy-Based Fine-Tuning of Language Models
- DeepRetrieval: Hacking Real Search Engines and Retrievers with Large Language Models via Reinforcement Learning
- SWE-bench: Can Language Models Resolve Real-World GitHub Issues?
- Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning
- Dense Passage Retrieval for Open-Domain Question Answering
- FAPO: Fully Automated Prompt Optimization of Multi-Step LLM Pipelines
- WebThinker: Empowering Large Reasoning Models with Deep Research Capability
- OpenResearcher: A Fully Open Pipeline for Long-Horizon Deep Research Trajectory Synthesis
- Synthesizing Multi-Agent Harnesses for Vulnerability Discovery
- The MiniMax-M2 Series: Mini Activations Unleashing Max Real-World Intelligence
- VulnLLM-R: Specialized Reasoning LLM with Agent Scaffold for Vulnerability Detection
- gpt-oss-120b & gpt-oss-20b Model Card
- TongSearch-QR: Reinforced Query Reasoning for Retrieval
- NYU CTF Bench: A Scalable Open-Source Benchmark Dataset for Evaluating LLMs in Offensive Security
- ReasonIR: Training Retrievers for Reasoning Tasks
Related papers
- SoK: AI-Augmented Binary Reversing
- Relaxed Sender Anonymity for CBDC Interbank Settlement: A Zero-Knowledge Approach on Permissioned EVM
- Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages
- Estimating the Social Cost of Corporate Data Breaches
- Efficient Fuzzy PSI under One-Sided Assumptions
- Sealing the Audit-Runtime Gap for LLM Skills