InCoder
Open-source code infilling model — generates code in the middle of existing functions, not just at the end.
Overview
A 2022 Meta AI research model (1.3B and 6.7B) that pioneered fill-in-the-middle code generation via causal masking, trained on 28 languages from GitHub and GitLab. Open weights on Hugging Face under a non-commercial licence; a research artifact, not a maintained product.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 11 Jul 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
Use cases
ASEAN Perspective
InCoder in Southeast Asia
ASEAN-region availability and pricing notes coming soon. Drop the editorial team a note via /contact/ if you can supply local context (Singapore/Malaysia/Indonesia/Thailand/Vietnam).
InCoder earns its place as a historical entry: it's the 2022 Meta AI research model that helped make fill-in-the-middle code generation standard practice, released as open 1.3B and 6.7B checkpoints on Hugging Face. Researchers studying infilling objectives or wanting a reproducible baseline will get value from it; working developers won't. It's long surpassed by Code Llama, StarCoder, DeepSeek-Coder and every commercial assistant, there's no hosted product or support, and the CC-BY-NC licence rules out commercial use. You bring your own inference. Fine for research and reproducibility; the wrong tool for day-to-day coding.
What people say
April 2022 is ancient history in code-model terms, and that's when InCoder landed — a Facebook AI Research project from Daniel Fried and colleagues that taught a transformer to fill in the middle of a function instead of only appending to the end. The trick was causal masking: mask a span during training, move it to the end, and the model learns to complete gaps using context on both sides. Two checkpoints shipped on Hugging Face, 1.3B and 6.7B parameters, trained on permissively licensed code from GitHub and GitLab across 28 languages plus StackOverflow text.
The idea won even though the model lost. Fill-in-the-middle is now table stakes — StarCoder, Code Llama and DeepSeek-Coder all train with some version of it, and every serious IDE assistant completes at the cursor rather than at the end of the file. InCoder itself, though, hasn't moved in years; Meta's code research went on to Code Llama in 2023 and the 32B Code World Model in September 2025.
Practical caveats: the weights carry a CC-BY-NC non-commercial licence, so you can't ship it in a product even if you wanted to. There's no hosted API, no support, and benchmark scores that modern 1B models beat handily. Load it if you're studying infilling objectives or need a citable 2022 baseline. For actual coding work, this belongs in the bibliography, not the toolchain.
Summary of public user & expert reviews, compiled by RECATOOLS.
Notable facts
- InCoder was the first large code model to demonstrate that left-to-right generation was not the optimal paradigm for real-world code editing tasks.
- The model achieved state-of-the-art fill-in-the-middle scores on the HumanEval infilling benchmark at its release in 2022.
- Facebook AI Research released InCoder months before GitHub Copilot's general availability, when the concept of AI code completion was still novel to most developers.
Frequently asked questions
About this listing
This entry was compiled from publicly available data including InCoder's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with InCoder unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to InCoder directly →
Spotted something out of date? Suggest an update →
InCoder in the news
Alternatives to InCoder
More in Code & Dev Tools
