# Codium-ai/AlphaCodium

Official implementation for the paper: "Code Generation with AlphaCodium: From Prompt Engineering to Flow Engineering""

Repository: https://github.com/Codium-ai/AlphaCodium
Canonical: https://ross.abutalabs.com/products/alphacodium
Homepage: https://www.codium.ai
Language: Python
License: AGPL-3.0
License Family: copyleft
Topics: code-generation, flow-engineering, paper-implementations, state-of-the-art, broader-impacts
Last push: 2024-11-25T13:09:34+00:00

## Health v2 (maintenance only)
Score: 26/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 68
- inputs: {"age_days": 962, "days_push": 646, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3965, forks 299 (observed 2026-08-28T04:08:30.670340+00:00)

## What it is
Official implementation of the AlphaCodium paper, a test-based, multi-stage, iterative flow for LLM code generation on competitive programming problems. It evaluates LLMs on the CodeContests dataset and significantly improves accuracy over single-prompt approaches.

## Use cases
- improve llm code generation accuracy on competitive programming problems
- reproduce the alphacodium paper results on codecontests
- benchmark gpt and claude models on code generation tasks
- learn flow engineering techniques for code-oriented llm pipelines
- evaluate pass@k accuracy of llms on codeforces-style problems

## When to choose
- you want state-of-the-art LLM accuracy on competitive programming or algorithmic code generation
- you are researching iterative, test-based code generation flows
- you need a reproducible benchmark harness for CodeContests

## When to avoid
- you need a production code review or code quality product rather than a research codebase
- you want general-purpose code completion inside an IDE
- you cannot provide an OpenAI API key or run Python-based evaluation pipelines

## Facets
- artifact type: library
- maturity: maintenance
- function: llm-inference, machine-learning, testing, prompt-engineering
- domain: large-language-models, artificial-intelligence, developer-tools
- platform: python, cli, cross-platform
- tags: code-generation, flow-engineering, competitive-programming, codecontests, paper-implementation, llm-evaluation, algorithms

## Member repositories
- Codium-ai/AlphaCodium (main) score 26

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:30.670340+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:24:24.816215+00:00, confidence not recorded.
  - readme: https://github.com/Codium-ai/AlphaCodium (fetched 2026-08-28T04:08:30.670340+00:00, sha 4c410914ba48)
  - homepage: https://www.codium.ai (fetched 2026-08-29T09:17:59.960904+00:00, sha ae35fa69a065)
  - site_page: https://www.qodo.ai/features/qodo-git (fetched 2026-08-29T09:17:59.963652+00:00, sha d40e4a17e168)
  - site_page: https://www.qodo.ai/features/qodo-context-engine (fetched 2026-08-29T09:17:59.965563+00:00, sha 894a9c4179b7)
  - site_page: https://docs.qodo.ai/qodo-documentation (fetched 2026-08-29T09:17:59.967254+00:00, sha 694b515b0265)
  - site_page: https://www.qodo.ai/about (fetched 2026-08-29T09:17:59.969783+00:00, sha c1c9cae225e1)
  - site_page: https://www.qodo.ai/pricing (fetched 2026-08-29T09:17:59.972246+00:00, sha c0ff8e8dda09)
  - site_page: https://www.qodo.ai/blog/qodo-integrations-turning-your-sdlc-into-review-context (fetched 2026-08-29T09:17:59.974767+00:00, sha a7389aa66a62)
- Data as of 2026-08-30T08:39:29.467469+00:00.