Run Llama 3.3 70B locally
Llama 3.3 70B is 70B parameters (dense). Here's the VRAM it needs at each quantization, the smallest GPU that fits, and how a self-hosted box compares with the cheapest verified API.
VRAM by quantization
| Quantization | VRAM needed | Fits on | Notes |
|---|---|---|---|
| int4 (Q4) | 42 GB | A100 (80 GB) ($1.19/hr) | smallest footprint, minor quality loss |
| int8 (Q8) | 84 GB | 2× H100 (160 GB) ($3.98/hr) | near-lossless |
| fp16 (full) | 168 GB | 4× H100 (320 GB) ($7.96/hr) | reference quality |
VRAM ≈ params × bytes/param × 1.2 (overhead). See the fullmethod and break-even calculator.
Self-host vs API
Cheapest API for Llama 3.3 70B is $0.42 / 1M tokens(blended) via DeepInfra. Running it yourself in int4 fits aA100 (80 GB) at $1.19/hr — about $868.7/month at 24/7. Those two lines cross at roughly 2.1B tokens/month: below that the API wins on cost, above it the dedicated GPU does (assuming you keep it busy). Tune your own volume in the break-even calculator.
Run it at home
No cloud account needed: Llama 3.3 70B in int4 (42 GB) fits a Mac Studio M5 Ultra (96 GB unified), a $5,499 one-off buy. Amortized over 3 years that's about $0.2092/hr — hardware only, electricity aside. Against the DeepInfra API at $0.42/1M, buying it pays for itself after roughly 13.1B tokens total. Below that the API is cheaper; a machine that mostly sits idle rarely earns back its price.
Get the weights
Quantized builds on the model hubs — links search live, so they track new builds as they appear:
- GGUF builds llama.cpp · Ollama · LM Studio
- AWQ builds GPU serving (vLLM)
- GPTQ builds GPU serving
- MLX builds Apple silicon
- Ollama library one-command local run
License: open weights under Llama Community License · terms. Compare all providers on the llama-3.3-70b API page.