Codestral-7B and Codestral-70B Throughput-Latency Trade-offs in Vulnerability Classification

This report synthesises findings from 2 peer-reviewed papers addressing the following research question: What is the throughput and latency trade-off between Codestral-7B and Codestral-70B when classifying vulnerabilities in Big-Vul under varying levels of parallelized inference and model quantization. 5 claims were extracted from source literature; 5 were independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 8.7/10. This report is a machine-generated literature synthesis and does not constitute original research.

Research goal: What is the throughput and latency trade-off between Codestral-7B and Codestral-70B when classifying vulnerabilities in Big-Vul under varying levels of parallelized inference and model quantization?

Autonomous literature synthesis. Automated review score: 8.7/10. Full text and citation available at Assignee Research.