Aporia’s 2024 Guardrails Benchmarks & multiSLM Detection Engine

Aporia has partnered with Portkey to bring secure and reliable Guardrails to production. The partnership aims to enhance the security and reliability of in-production GenAI applications for businesses. Aporia's Guardrails have undergone rigorous benchmarking, achieving an average latency of 0.34 seconds and a 90th percentile latency of 0.43 seconds. The company outperforms NeMo, GPT-4o, and GPT-3.5 in both hallucination detection accuracy and latency. Aporia's multiSLM Detection Engine employs multiple smaller language models (SLMs) instead of a single large language model (LLM), ensuring greater robustness and flexibility in the Guardrails and policies offered. The use of SLMs over 1 big LLM has allowed Aporia to create lightning-fast and highly accurate Guardrails that set new standards in the industry, benefiting users by securing their AI from malicious threats or inaccurate responses.