CodeLlama
Meta's open-source code LLM family — specialised versions for Python, instruction following, and code infill.
Overview
Code Llama is Meta's open-source language model family specifically fine-tuned for coding tasks, built on top of the base Llama 2 and Llama 3 models. Released in August 2023, it comes in three variants: base Code Llama (general coding), Code Llama - Python (optimised specifically for Python), and Code Llama - Instruct (follows natural language instructions for code generation).
Code Llama supports code infill — given existing code with a gap, the model can fill in the middle portion rather than just generating from the beginning. This is essential for use cases like autocomplete in the middle of a function. The model handles context windows up to 100,000 tokens, allowing analysis of very large code files.
Available in 7B, 13B, 34B, and 70B parameter sizes, Code Llama provides a range of quality/compute tradeoffs. The 7B model runs efficiently on consumer hardware for local coding assistance; the 70B model approaches the quality of commercial alternatives. The permissive licence (Llama Community Licence) allows commercial use for most organisations.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 8 May 2026 and may be out of date or incomplete.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
CodeLlama 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).
Code Llama is Meta's open-weight family of code-specialised LLMs (7B to 70B, plus Python and Instruct variants) built on Llama 2, supporting completion, infilling and instruction following. As a free, self-hostable foundation it was a landmark release and remains a reasonable base for teams building private coding tools, though newer open models (and Meta's own later releases) have since surpassed it.
It suits ML engineers, researchers and privacy-conscious teams wanting to run or fine-tune a code model locally, not end users wanting a ready assistant. Honest caveats: it is weights, not a product, so you supply the serving, tooling and UX; raw quality now lags current open and commercial frontier coders; and the Llama community licence has usage conditions worth checking. No hosted API from Meta; you self-host or use third-party providers. ASEAN teams can deploy it anywhere, which is a real sovereignty/cost advantage.
Notable facts
- Code Llama supports code infill at 100,000 token context — long enough to fit over 2,000 lines of Python in a single context window.
- The Python-specialised variant was trained with a much higher proportion of Python code than the base model, outperforming the base on Python-specific benchmarks.
- Code Llama 70B was the first openly available model to pass 50% on the HumanEval coding benchmark, a major milestone for open-source code models.
Frequently asked questions
About this listing
This entry was compiled from publicly available data including CodeLlama's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with CodeLlama 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 CodeLlama directly →
Spotted something out of date? Suggest an update →
Alternatives to CodeLlama
More in Code & Dev Tools
