Model · Meta · Text
Llama 3.3 70B
Meta's latest 70B-parameter open-weight model. Reaches frontier-tier performance for English-centric tasks while remaining self-hostable.
- Modality
- Text
- License
- Llama 3 Community License (Open weights)
- Parameter size
- 70B
- Context window
- 128,000 tokens
- Released
- December 6, 2024
- Last verified
- June 8, 2026
- Runs locally
- Yes
What you need to run it
- Mac
- Yes. Any Apple Silicon Mac (M1 or newer) with at least 48 GB (64 GB recommended) of memory. Newer chips are faster, but memory is what decides whether it runs.
- Windows / Linux PC
- NVIDIA works best. A graphics card with at least 48 GB (80 GB recommended) of video memory.
- No graphics card
- Not practical. Too slow without a graphics card or Apple Silicon.
Runs on a 48 GB or 64 GB Mac in Ollama, LM Studio or MLX, though it generates text fairly slowly because every part of it is used for each word.
Numbers are for: 4-bit build of the 70B model (about 43 GB). Our estimate from the model’s size — the maker publishes no figure.
Strengths
- Open weights — runs offline on capable hardware
- 128k context window
- Strong instruction following for an open model
Weaknesses
- Needs a multi-GPU setup or quantization for local use
- Community License restricts use at >700M MAU
Try it
| Where | Type | Notes |
|---|---|---|
| Hugging Face | weights | Gated — accept license to download |
| OpenRouter | hosted-api | API key required |
| Groq | hosted-api | Free tier, very fast inference |
| Ollama | local | ollama run llama3.3 |
Used in solutions
Version history
- Llama 3.3 70B Dec 2024 Current
- Llama 3.2 Vision Sep 2024
- Llama 3.1 8B Jul 2024
Official sources
- Model card model
Change log
- — Added hardware requirements (Mac support, memory, NVIDIA). Source: estimate from 70B size at 4-bit.
- — Initial entry.
Esc