Skip to content

Kanti Labs

Kanti Labs Models

Kanti Labs models are open-weight offensive security models built for web exploitation research, local deployment, and iterative post-training.

SFT

kanti-xss-sft-4b

Supervised fine-tuned on curated XSS synthetic Strix traces from real vulnerabilities. Designed to be used as a Strix sub-agent.

BaseQwen3-4B
MethodSupervised Fine-Tuning
TaskXSS Payload Generation
RL

kanti-xss-rl-4b

Reinforcement learning variant trained with reward signals from simulated XSS validation. Designed to be used as a Strix sub-agent.

BaseQwen3-4B
MethodReinforcement Learning
TaskXSS Payload Generation

Open Weight

Kanti Labs models are released so researchers can inspect weights, reproduce local runs, and test real integration constraints instead of relying on a closed demo stack.

Web Exploitation Focus

The current series concentrates on XSS workflows, synthetic traces, and reward signals that map directly to offensive web security tasks.

Built for Iteration

Model cards are only part of the picture. Kanti Labs pairs releases with research notes and a public roadmap so the training direction stays legible.

Related Kanti Labs Pages

Read the Kanti Labs research roadmap for training context, or use Contact Kanti Labs to discuss collaboration or model evaluation.