LLaMA-33B
LLaMA-33B is an open-weights language model from Meta, released February 24, 2023. It has 33B parameters (dense). On the Epoch Capabilities Index it scores 96.6, rank 222 of 256 rated models.
Check it against your own PC →
Can you run LLaMA-33B locally?
The smallest common GPU it runs on is 16 GB of VRAM (with 32 GB of RAM): IQ3_XXS, split GPU / CPU, roughly 11 tokens per second. Estimates assume 32 GB of DDR5 RAM, a 16K context and llama.cpp-style GGUF quantization. Enter your exact hardware in the live app.
| GPU memory | Verdict | Best quant | VRAM | RAM | tok/s |
|---|---|---|---|---|---|
| 6 GB VRAM | Very slow | IQ3_XXS (noticeable loss) | 5.2 GB | 11.4 GB | ~3 |
| 8 GB VRAM | Very slow | IQ3_XXS (noticeable loss) | 7.2 GB | 9.4 GB | ~3 |
| 12 GB VRAM | Very slow | IQ3_XXS (noticeable loss) | 11.2 GB | 5.4 GB | ~5 |
| 16 GB VRAM | Runs, slowly | IQ3_XXS (noticeable loss) | 15.2 GB | 1.4 GB | ~11 |
| 24 GB VRAM | Runs on GPU | IQ4_XS (good) | 21.5 GB | – | ~31 |
| 32 GB VRAM | Runs on GPU | Q6_K (≈ lossless) | 31.1 GB | – | ~38 |
Benchmark results
| Benchmark (Epoch AI) | Score |
|---|---|
| MMLU | 44.9% |
Links
Models with a similar score
DeepSeek-Coder-V2-Lite-Base · Falcon 2 11B · Nemotron-4 15B · internlm-20b · PaLM 2-M · Qwen2.5-Coder-3B · Llama 2-13B · Phi-2
More from Meta
Muse Ember · Muse Spark Open · Muse Wonder · Muse Wonder Full · Muse Spark 1.3 · Muse Spark 1.3 Contributor · Muse Glimmer · Muse Glimmer 30B assistant