NVIDIA · 2020-11
Best local LLMs for the A100 80 GB
80 GB VRAM · 2039 GB/s · ~79 GB usable.
Check A100 80 GB price →Try it with your context & use case
Preset to the A100 80 GB. Change the context length or filter by use case.
28 models run well on A100 80 GB (any context — just fitting the weights) — 79 GB usable.
Sort:
| Model | Size | Fit | Best quant | Needs | Memory | Speed |
|---|---|---|---|---|---|---|
| Llama 4 Scout (MoE)MoE | 109B | Runs well | Q5_K_M | 76 GB | ~79 tok/s | |
| Qwen2.5 72B | 72.7B | Runs well | Q8_0 | 76 GB | ~19 tok/s | |
| Llama 3.3 70B | 70.6B | Runs well | Q8_0 | 73 GB | ~20 tok/s | |
| DeepSeek-R1 Distill Llama 70B | 70.6B | Runs well | Q8_0 | 73 GB | ~20 tok/s | |
| Mixtral 8x7B (MoE)MoE | 46.7B | Runs well | Q8_0 | 49 GB | ~70 tok/s | |
| Qwen3 32B | 32.8B | Runs well | FP16 | 64 GB | ~22 tok/s | |
| Qwen2.5 32B | 32.8B | Runs well | FP16 | 64 GB | ~22 tok/s | |
| Qwen2.5-Coder 32B | 32.8B | Runs well | FP16 | 64 GB | ~22 tok/s | |
| DeepSeek-R1 Distill Qwen 32B | 32.8B | Runs well | FP16 | 64 GB | ~22 tok/s | |
| Qwen3 30B-A3B (MoE)MoE | 30.5B | Runs well | FP16 | 60 GB | ~145 tok/s | |
| Gemma 3 27B | 27.4B | Runs well | FP16 | 54 GB | ~27 tok/s | |
| Gemma 2 27B | 27.2B | Runs well | FP16 | 53 GB | ~27 tok/s | |
| Mistral Small 3 24B | 23.6B | Runs well | FP16 | 46 GB | ~31 tok/s | |
| Qwen3 14B | 14.8B | Runs well | FP16 | 29 GB | ~50 tok/s | |
| Qwen2.5 14B | 14.8B | Runs well | FP16 | 29 GB | ~50 tok/s | |
| DeepSeek-R1 Distill Qwen 14B | 14.8B | Runs well | FP16 | 29 GB | ~50 tok/s | |
| Phi-4 14B | 14.7B | Runs well | FP16 | 29 GB | ~50 tok/s | |
| Mistral Nemo 12B | 12.2B | Runs well | FP16 | 24 GB | ~60 tok/s | |
| Gemma 3 12B | 12.2B | Runs well | FP16 | 24 GB | ~60 tok/s | |
| Gemma 2 9B | 9.2B | Runs well | FP16 | 19 GB | ~79 tok/s | |
| Qwen3 8B | 8.2B | Runs well | FP16 | 17 GB | ~90 tok/s | |
| Llama 3.1 8B | 8.0B | Runs well | FP16 | 16 GB | ~91 tok/s | |
| Qwen2.5-Coder 7B | 7.6B | Runs well | FP16 | 15 GB | ~97 tok/s | |
| Mistral 7B v0.3 | 7.3B | Runs well | FP16 | 15 GB | ~101 tok/s | |
| Gemma 3 4B | 4.3B | Runs well | FP16 | 9.1 GB | ~171 tok/s | |
| Qwen3 4B | 4B | Runs well | FP16 | 8.5 GB | ~184 tok/s | |
| Llama 3.2 3B | 3.2B | Runs well | FP16 | 7.0 GB | ~229 tok/s | |
| Llama 3.2 1B | 1.2B | Runs well | FP16 | 3.2 GB | ~592 tok/s | |
| DeepSeek-R1 671B-A37B (MoE)MoE | 671B | Won't fit | — | 273 GB | — | |
| Qwen3 235B-A22B (MoE)MoE | 235B | Won't fit | — | 96 GB | — |
Estimates, computed in your browser — VRAM, quantisation and speed vary with your runtime and settings. How we estimate →
Top picks for the A100 80 GB
Every model on the A100 80 GB
| Model | Size | Fit | Best quant | Needs | Speed |
|---|---|---|---|---|---|
| Llama 4 Scout (MoE)MoE | 109B | Runs well | Q5 | 77 GB | ~79 tok/s |
| Qwen2.5 72B | 72.7B | Runs well | Q8 | 78 GB | ~19 tok/s |
| Llama 3.3 70B | 70.6B | Runs well | Q8 | 76 GB | ~20 tok/s |
| DeepSeek-R1 Distill Llama 70B | 70.6B | Runs well | Q8 | 76 GB | ~20 tok/s |
| Mixtral 8x7B (MoE)MoE | 46.7B | Runs well | Q8 | 50 GB | ~70 tok/s |
| Qwen3 32B | 32.8B | Runs well | FP16 | 66 GB | ~22 tok/s |
| Qwen2.5 32B | 32.8B | Runs well | FP16 | 66 GB | ~22 tok/s |
| Qwen2.5-Coder 32B | 32.8B | Runs well | FP16 | 66 GB | ~22 tok/s |
| DeepSeek-R1 Distill Qwen 32B | 32.8B | Runs well | FP16 | 66 GB | ~22 tok/s |
| Qwen3 30B-A3B (MoE)MoE | 30.5B | Runs well | FP16 | 61 GB | ~145 tok/s |
| Gemma 3 27B | 27.4B | Runs well | FP16 | 58 GB | ~27 tok/s |
| Gemma 2 27B | 27.2B | Runs well | FP16 | 56 GB | ~27 tok/s |
| Mistral Small 3 24B | 23.6B | Runs well | FP16 | 48 GB | ~31 tok/s |
| Qwen3 14B | 14.8B | Runs well | FP16 | 31 GB | ~50 tok/s |
| Qwen2.5 14B | 14.8B | Runs well | FP16 | 31 GB | ~50 tok/s |
| DeepSeek-R1 Distill Qwen 14B | 14.8B | Runs well | FP16 | 31 GB | ~50 tok/s |
| Phi-4 14B | 14.7B | Runs well | FP16 | 31 GB | ~50 tok/s |
| Mistral Nemo 12B | 12.2B | Runs well | FP16 | 26 GB | ~60 tok/s |
| Gemma 3 12B | 12.2B | Runs well | FP16 | 27 GB | ~60 tok/s |
| Gemma 2 9B | 9.2B | Runs well | FP16 | 21 GB | ~79 tok/s |
| Qwen3 8B | 8.2B | Runs well | FP16 | 18 GB | ~90 tok/s |
| Llama 3.1 8B | 8.0B | Runs well | FP16 | 17 GB | ~91 tok/s |
| Qwen2.5-Coder 7B | 7.6B | Runs well | FP16 | 16 GB | ~97 tok/s |
| Mistral 7B v0.3 | 7.3B | Runs well | FP16 | 16 GB | ~101 tok/s |
| Gemma 3 4B | 4.3B | Runs well | FP16 | 10 GB | ~171 tok/s |
| Qwen3 4B | 4B | Runs well | FP16 | 9.6 GB | ~184 tok/s |
| Llama 3.2 3B | 3.2B | Runs well | FP16 | 7.8 GB | ~229 tok/s |
| Llama 3.2 1B | 1.2B | Runs well | FP16 | 3.4 GB | ~592 tok/s |
| DeepSeek-R1 671B-A37B (MoE)MoE | 671B | Won't fit | — | 303 GB | — |
| Qwen3 235B-A22B (MoE)MoE | 235B | Won't fit | — | 98 GB | — |
Similar hardware
FAQ
What is the best LLM for the A100 80 GB?
For general use, Llama 4 Scout (MoE) is the strongest model that runs well on the A100 80 GB. See the picks-by-use-case below for coding, reasoning and more.
How much can the A100 80 GB run?
The A100 80 GB has 80 GB of VRAM, of which about 79 GB is usable for a model. That runs 28 of our tracked models well.
Estimates at 8K context — see how we compute these.