Skip to content
Last Dictionary
Plane
Language Models & TransformersAgents & HarnessesAI Engineering & MLOps

code generation

nounAlso: codegen, AI code generation, code synthesis, text-to-code, code completion

In plain wordsHaving an AI write computer code from a description in plain language, from finishing a single line as you type to building whole programs.

  1. 1

    AI

    The automatic production of source code by a model from a natural-language specification, surrounding code, tests or examples. Modes range from inline completion and fill-in-the-middlefill-in-the-middle, n.: A training trick that teaches a text-writing AI to fill a gap using what comes both before and after it — essential for coding assistants that insert code in the middle of a file. (predicting code between a given prefix and suffix) through function-level synthesis from docstrings to repository-wide edits and agentic codingagentic coding, n.: Writing software by directing AI agents that plan, edit and test code themselves, with the human acting more as a reviewer and manager than as the person typing every line., in which a coding agentcoding agent, n.: An AI agent that writes and fixes software by itself: it reads a codebase, edits files, runs commands and tests, and keeps working until a feature is built or a bug is fixed. writes, runs and debugs code with tools. Code models are trained on large public code corpora, then post-trained with execution feedback and verifiable rewardsreinforcement learning with verifiable rewards, n.: Training an AI by letting it attempt problems whose answers can be checked automatically — a math result, a passing program — and rewarding it only when the check succeeds., since tests make correctness checkable. Benchmarks progressed from HumanEvalHumanEval, prop. n.: A set of 164 small programming exercises, each described in plain English and checked by hidden tests, used to see whether an AI can write working Python code. (function completion, scored by pass@kpass@k, n.: A score for code-writing AI: the chance that at least one of k attempts at a problem is correct, as checked by actually running tests on the code.) to real-repository tasks such as SWE-benchSWE-bench, prop. n.: A test that asks AI to fix real bugs and add features in real open-source Python projects, checked by running each project's own tests; the best-known yardstick for AI coding agents.. Risks include subtly insecure or license-encumbered code, invented package names (slopsquattingslopsquatting, n.: An attack that exploits AI coding assistants' habit of inventing software packages that do not exist: attackers publish malicious code under those made-up names and wait for people or agents to install it.) and erosion of developers' skills.

    Code generation covers our boilerplate; humans still review anything touching payments.

  2. 2

    Compilers

    The compiler phase that translates a program's intermediate representation into target machine code or bytecode, including instruction selection, register allocation and scheduling. This older sense is unrelated to machine learning, except where AI compilersAI compiler, n.: Software that translates an AI model into efficient instructions for a specific chip, rearranging and combining its calculations so the hardware is used as fully as possible. generate kernels for accelerators.

Origin

The compiler sense dates from the first optimizing compilers of the 1950s. Neural code generation from natural language was studied from about 2016; OpenAI's CodexCodex, prop. n.: A name OpenAI has used twice: first for a 2021 AI model that wrote code and powered the original GitHub Copilot, then from 2025 for its family of coding agents. paper (Mark Chen and colleagues, July 2021), which introduced HumanEval, and GitHub CopilotGitHub Copilot, prop. n.: An AI assistant from GitHub, built into programmers' code editors, that suggests code as they type, answers questions about code and can carry out coding tasks on its own.'s preview the same summer made LLM code generation mainstream, followed by DeepMind's AlphaCode (2022).

Related

See also
  • coding agentcoding agent, n.: An AI agent that writes and fixes software by itself: it reads a codebase, edits files, runs commands and tests, and keeps working until a feature is built or a bug is fixed.
  • agentic codingagentic coding, n.: Writing software by directing AI agents that plan, edit and test code themselves, with the human acting more as a reviewer and manager than as the person typing every line.
  • program synthesisprogram synthesis, n.: Getting a computer to write a program automatically from a description of what the program should do — a logical specification, a few input–output examples or a request in plain language.
  • HumanEvalHumanEval, prop. n.: A set of 164 small programming exercises, each described in plain English and checked by hidden tests, used to see whether an AI can write working Python code.
  • SWE-benchSWE-bench, prop. n.: A test that asks AI to fix real bugs and add features in real open-source Python projects, checked by running each project's own tests; the best-known yardstick for AI coding agents.
  • GitHub CopilotGitHub Copilot, prop. n.: An AI assistant from GitHub, built into programmers' code editors, that suggests code as they type, answers questions about code and can carry out coding tasks on its own.
  • vibe codingvibe coding, n.: Programming by conversation: you ask an AI for what you want, try out what it produces and ask again until it works, trusting the code without really reading it.
  • CodexCodex, prop. n.: A name OpenAI has used twice: first for a 2021 AI model that wrote code and powered the original GitHub Copilot, then from 2025 for its family of coding agents.

Sources

  1. Chen, M. et al. (2021). Evaluating Large Language Models Trained on Code. arXiv:2107.03374.

Mentioned in

coding agentcoding agent, n.: An AI agent that writes and fixes software by itself: it reads a codebase, edits files, runs commands and tests, and keeps working until a feature is built or a bug is fixed., GitHub CopilotGitHub Copilot, prop. n.: An AI assistant from GitHub, built into programmers' code editors, that suggests code as they type, answers questions about code and can carry out coding tasks on its own., AI coding assistantAI coding assistant, n.: AI software that helps people write computer programs — suggesting the next lines, explaining code, finding bugs or making changes on request — inside the tools programmers already use., code executioncode execution, n.: Letting an AI write a small program and actually run it — to do math, analyze a spreadsheet or test its own code — then use the real output instead of guessing., completioncompletion, n.: The text an AI language model writes in response to a prompt — literally, its continuation of the text it was given., HumanEvalHumanEval, prop. n.: A set of 164 small programming exercises, each described in plain English and checked by hidden tests, used to see whether an AI can write working Python code., pass@kpass@k, n.: A score for code-writing AI: the chance that at least one of k attempts at a problem is correct, as checked by actually running tests on the code., program synthesisprogram synthesis, n.: Getting a computer to write a program automatically from a description of what the program should do — a logical specification, a few input–output examples or a request in plain language., CursorCursor, prop. n.: A code editor built around AI, in which programmers can ask a model to write, change or explain code across a whole project, or hand it tasks to finish on its own., fill-in-the-middlefill-in-the-middle, n.: A training trick that teaches a text-writing AI to fill a gap using what comes both before and after it — essential for coding assistants that insert code in the middle of a file., program-aided language modelprogram-aided language model, n.: A technique in which an AI solves a math or logic problem by writing a small computer program and letting a computer run it, instead of doing the arithmetic in its head., semantic parsingsemantic parsing, n.: Translating a sentence into a precise, machine-executable form — such as a database query or a logical formula — so that a computer can act on what was asked., The StackThe Stack, prop. n.: A large free collection of computer source code from public repositories, selected mainly for permissive licenses and released from 2022 onward by the BigCode project for training coding models.