Can I run this LLM?
Can a NVIDIA RTX 3090 (24GB) run Llama 3.2 3B?
Llama · Llama 3.2 Community · estimated for a 24.0 GB machine at 4K context.
Yes — it runs great.
Runs greatBest fit: Llama 3.2 3B at FP16 · 6.4 GB weights · 79–118 tok/s on a NVIDIA RTX 3090 (24GB).
Arena Text Overall · style-controlled
1,167FP16 untestedllama-3.2-3b-instruct · published configuration reference · FP16 not tested
Arena Text Overall (style-controlled), exact published entry llama-3.2-3b-instruct, 2026-10-08: reference rating 1,166.8. Published confidence bounds 1,159.1–1,174.5. 7,936 votes. Published llama-3.2-3b-instruct configuration reference. Reasoning mode is not disclosed by the Arena dataset. Tested precision, runtime and sampling settings are not disclosed in this dataset. Selected FP16 has not been measured by this Arena entry. No quantization quality adjustment is applied.
Snapshot published 2026-10-08 · snapshot imported 2026-10-11 UTC · validated last-good data · manual refresh
Arena leaderboard dataset · Arena / lmarena-ai, Arena Leaderboard Dataset (CC BY 4.0). Selected Text Overall style-controlled rows; field names adapted, ratings unchanged. · dataset CC BY 4.0 · revision e16f90232986
- Total memory at 4K context
- 7.7 GB
- Weights + KV cache + runtime overhead
- Estimated generation
- 79–118 tok/s
- Estimated first token
- ~1.2s
Every quantization on this machine
On small screens, each quant lists every metric below its name. On wider screens, scroll sideways for total memory, speed and fit. Keyboard: focus the table, then use the arrow keys. Quant names stay visible.
| Quant | Weights | Total memory | Speed | Fit |
|---|---|---|---|---|
| Q4_K_M | 2.0 GB | 3.3 GB | 158–237 tok/s | Runs great |
| Q5_K_M | 2.3 GB | 3.6 GB | 148–222 tok/s | Runs great |
| Q6_K | 2.6 GB | 3.9 GB | 138–208 tok/s | Runs great |
| Q8_0 | 3.4 GB | 4.7 GB | 120–180 tok/s | Runs great |
| FP16 | 6.4 GB | 7.7 GB | 79–118 tok/s | Runs great |
Ran Llama 3.2 3B on a NVIDIA RTX 3090 (24GB)? Add your real numbers →