Skip to content
← Models
Open weightsMeta/10 trustMetrics-checked
CodeLlama 34B logo

CodeLlama 34B

Meta's pioneering code LLM — the model that proved open code models could compete.

Visit official page →

About

CodeLlama-34B-Instruct (August 2023) is Meta's pioneering code LLM: the model that proved open-weights code models could compete with the closed leaders of 2023, spawning the entire open code-model ecosystem that followed.

Licensed under the Llama 2 Community License: custom, NOT open source — commercial use allowed under the terms but with restrictions (including the 700M MAU clause). Review before product use. Historically vital, currently obsolete for new builds: 2023-era quality, 16K context, and superseded many times over (Qwen2.5-Coder-7B beats it at 1/5 the size). Direct adoption is minimal (397 HF downloads on the canonical checkpoint). Included as the historical landmark of open code models — choose Qwen2.5-Coder for actual work.

Specs

Context window

16,384 tokens

Pricing

Free to download; commercial use under Llama 2 Community License (custom terms, 700M MAU clause).

Free quota

Free — download and run the weights yourself; review Llama 2 license terms.

API access

Self-hosted (Transformers; Hugging Face checkpoints).

Weights

Open weights

Released

August 24, 2023

Benchmarks

Provider-claimed at release (2023): leading open HumanEval — treat as provider-claimed and dated. Superseded many times over.

Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.

Use cases

codecodingresearchhistory

Trust score — 4/10

/10 trustMetrics-checked

Scored 4/10 (metrics tier). Historically the most important open code model — it started the ecosystem. But this is 2023: (1) Llama 2 Community License is custom, not open source; (2) quality is obsolete — Qwen2.5-Coder-7B beats it at a fifth the size; (3) 16K context; (4) minimal current adoption (397 HF downloads). A museum piece: respect it, don't deploy it.

How the trust score is built

Full methodology →
Adoption · stars/downloads, log-scaled vs peers
4.4/10
Maintenance · recency, releases, issue load
4.0/10
Openness · license + docs
4.6/10
Manual review · sanity read + red-flag check
4.0/10
  • · HF downloads 397 (peer-normalized)
  • · released 1146 days ago
  • · manual review: historically vital, currently obsolete; Qwen2.5-Coder-7B beats it at 1/5 size; museum piece; license: Llama 2 Community License — custom, not open source (700M MAU clause)

More in the codellama family

Sources