Llama 3.3 — local AI model by Meta
Written by Jakub Rusinowski · Last updated
Meta's latest flagship model. Llama 3.3 70B delivers state-of-the-art performance, matching the massive 405B model in many benchmarks while remaining efficient enough for high-end local setups.
Variants
The smallest Llama 3.3 variant needs about 43 GB of VRAM at Q2_K_XS (Tight) — quantized weights plus framework overhead, before any KV cache.
| Model | VRAM at Q4 | VRAM | Context | Run it |
|---|---|---|---|---|
| Llama 3.3 70B Instruct → 70B | ~43.1 GB | 128,000 | ollama run llama3.3 |
Memory is quantized weights plus overhead at Q4_K_M, from the same engine as the GPU & VRAM checker.
How to run Llama 3.3 locally
Install Ollama, then pull the tag.
ollama run llama3.3Pick a size above for its own VRAM figure, speed estimate and install command.
Licence
Weights are downloadable and commercial use is permitted, subject to the licence’s acceptable-use terms.
Applies to: Llama 3.3 70B InstructRecommended GPU
The cheapest catalogued GPU that runs Llama 3.3 locally (min 43 GB VRAM) is the NVIDIA DGX Spark (128 GB).
Can I run Llama 3.3 on my GPU?
- Llama 3.3 on AMD Radeon RX 7800 XT
- Llama 3.3 on AMD Radeon RX 7900 XT
- Llama 3.3 on AMD Radeon RX 7900 XTX
- Llama 3.3 on AMD Radeon RX 9060 XT 16GB
- Llama 3.3 on AMD Radeon RX 9070
- Llama 3.3 on AMD Radeon RX 9070 XT
- Llama 3.3 on Apple M1
- Llama 3.3 on Apple M1 Pro
- Llama 3.3 on Apple M2
- Llama 3.3 on Apple M2 Pro
- Llama 3.3 on Apple M3
- Llama 3.3 on Apple M3 Pro
- Llama 3.3 on Apple M4
- Llama 3.3 on Apple M4 Pro
- Llama 3.3 on Apple M5
- Llama 3.3 on Intel Arc B570
- Llama 3.3 on Intel Arc B580
- Llama 3.3 on NVIDIA GeForce RTX 3060 (12GB)
- Llama 3.3 on NVIDIA GeForce RTX 3080 (10GB)
- Llama 3.3 on NVIDIA GeForce RTX 3090
- Llama 3.3 on NVIDIA GeForce RTX 4060 Ti 16GB
- Llama 3.3 on NVIDIA GeForce RTX 4070
- Llama 3.3 on NVIDIA GeForce RTX 4070 Super
- Llama 3.3 on NVIDIA GeForce RTX 4070 Ti
- Llama 3.3 on NVIDIA GeForce RTX 4070 Ti Super
- Llama 3.3 on NVIDIA GeForce RTX 4080
- Llama 3.3 on NVIDIA GeForce RTX 4080 Super
- Llama 3.3 on NVIDIA GeForce RTX 4090
- Llama 3.3 on NVIDIA GeForce RTX 5060 Ti 16GB
- Llama 3.3 on NVIDIA GeForce RTX 5070
- Llama 3.3 on NVIDIA GeForce RTX 5070 Ti
- Llama 3.3 on NVIDIA GeForce RTX 5080
- Llama 3.3 on NVIDIA GeForce RTX 5090
- Llama 3.3 on NVIDIA L40S
- Llama 3.3 on NVIDIA RTX 6000 Ada Generation
Llama 3.3 — frequently asked questions
How much VRAM does Llama 3.3 need?
Llama 3.3 needs about 43 GB VRAM at Q2_K_XS (Tight) quantization for its smallest variant. Variants: Llama 3.3 70B Instruct (43 GB, Q2_K_XS (Tight)). On Apple Silicon, unified memory counts toward this requirement.
Can I run Llama 3.3 on an RTX 4090 (24 GB)?
Llama 3.3's smallest variant needs about 43 GB, which exceeds a single RTX 4090 (24 GB). Use multiple GPUs, a higher-VRAM card, or Apple Silicon with large unified memory.
What quantization should I use for Llama 3.3?
Q4_K_M is the best balance of quality and VRAM for Llama 3.3 in most cases. Choose Q8_0 for near-lossless quality if you have spare VRAM, or smaller quants (Q3/Q2) only when memory is tight.
How do I run Llama 3.3 with Ollama?
Install Ollama, then run: ollama run llama3.3. This downloads Llama 3.3 and starts a local, OpenAI-compatible endpoint — no internet connection is needed after the initial download.