Model Size and Inference Efficiency Trade-offs in Distilled Code Generation Models

This report synthesises findings from 12 peer-reviewed papers addressing the following research question: How does the trade-off between model size and inference efficiency vary when distilling code generation capabilities from large language models to smaller models, as measured by latency and pass@k. Abstract The rapid evolution of large language models (LLMs) has driven a transformative shift in artificial intelligence (AI), reshaping both research paradigms and practical applications. Distinguished from their predecessors by unprecedented scale and advanced capabilities. 9 claims were extracted from source literature; 9 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: How does the trade-off between model size and inference efficiency vary when distilling code generation capabilities from large language models to smaller models, as measured by latency and pass@k metrics across different programming languages on HumanEval and MBPP benchmarks?

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