This browser is not keeping saved items, so they will be forgotten when you leave this page. Everything else still works.
Reference · Open-weight model family
Llama
11 facts collected from their sources, each linked with its date. Nothing here is written from memory; a fact appears only as its source states it.
Models on Hugging Face
- meta-llama/Llama-Prompt-Guard-2-86M is listed on Hugging Face (text-classification, transformers, license:other). Meta Llama on Hugging Face,
- meta-llama/Llama-Prompt-Guard-2-22M is listed on Hugging Face (text-classification, transformers, license:other). Meta Llama on Hugging Face,
- meta-llama/Llama-Guard-4-12B is listed on Hugging Face (image-text-to-text, transformers, license:other). Meta Llama on Hugging Face,
- meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-Original is listed on Hugging Face (license:other). Meta Llama on Hugging Face,
- meta-llama/Llama-4-Maverick-17B-128E-Instruct-Original is listed on Hugging Face (license:other). Meta Llama on Hugging Face,
- meta-llama/Llama-4-Maverick-17B-128E-Original is listed on Hugging Face (license:other). Meta Llama on Hugging Face,
Benchmark results
- mini-SWE-agent with llama-4-maverick-instruct: 21.04% (resolved) on SWE-bench, Verified. SWE-bench leaderboard,
- mini-SWE-agent with llama-4-scout-instruct: 9.06% (resolved) on SWE-bench, Verified. SWE-bench leaderboard,
- Llama 4 Maverick: 15.6% (pass rate, second attempt) on Aider polyglot, 225 exercises. Aider polyglot leaderboard,
- RAG baseline with SWE-Llama 13B: 1.2% (resolved) on SWE-bench, Verified. SWE-bench leaderboard,
- RAG baseline with SWE-Llama: 1.4% (resolved) on SWE-bench, Verified. SWE-bench leaderboard,
Related on aibrief
- Stories
- Comparisons