Overview

Code generation asks how models can turn requests, examples, images, or formal specifications into programs that work in real settings, from short functions to whole repositories and domain-specific software. Common approaches include prompting and fine-tuning, retrieval of relevant code or documentation, constrained decoding, and agents that iteratively run, test, and repair their outputs. Multimodal systems also use rendered interfaces, charts, and diagrams as input or feedback, while reinforcement learning can use execution tests, compilers, or formal verifiers as training signals. Benchmarks such as HumanEval, MBPP, LiveCodeBench, and repository-level tasks show progress, but passing tests or compiling does not guarantee that code meets its intended behavior, visual requirements, security needs, or deployment constraints.

Latest papers

CardsList

335 papers34 papers in the last four weeks, level with the four weeks before. 0.2% of all new papers.

  1. BigO(Bench): Can LLMs Generate Code with Controlled Time and Space Complexity?

    Mar 19, 2025Pierre Chambon, Baptiste Roziere, Benoit Sagot +1Benchmark DesignCode Generation

  2. Leveraging Metamemory Agent for Enhanced Data-Free Code Generation in Large Language Models

    Jan 14, 2025Shengsheng Zhou, Shuai Wang, Liang Ding +5Code GenerationAgentic Code Generation

  3. BigCodeBench: Benchmarking Code Generation with Diverse Function Calls and Complex Instructions

    Jun 22, 2024Terry Yue Zhuo, Minh Chien Vu, Jenny Chim +30Code Generation BenchmarksCode Generation

  4. Astraios: Parameter-Efficient Instruction Tuning Code Large Language Models

    Jan 1, 2024Terry Yue Zhuo, Armel Zebaze, Nitchakarn Suppattarachai +4LLM Fine-TuningInstruction Tuning