ToolSite

What LLMs can an A100 80GB run?

An A100 80GB gives a model 80 GB of memory and 2,039 GB/s of bandwidth. Of the 40 open models tracked here, 23 fit at Q4_K_M with an 8,192-token context, and 2 more at a lower precision.

Models that fit on one A100 80GB

The most precise weights that still fit with 8K tokens of context and one request, the memory that takes, the longest context at that precision and the writing speed. Bigger models come first.

Model Best precision Memory Longest context Tokens/s
Qwen3.8 Flash Next 180.0B, 6B active GGUF Q2_K 77.9 GB 88K 168–316
Mistral Medium 3.5 128B 127.7B GGUF Q3_K_M 67.5 GB 41K 17–23
Qwen3.5 122B-A10B 125.1B, 10B active GGUF Q4_K_M 78.2 GB 76K 85–151
Nemotron 3 Super 120B-A12B 123.6B, 12B active GGUF Q4_K_M 77.2 GB 256K (full) 74–130
gpt-oss-120b 116.8B, 5.1B active GGUF Q4_K_M 73.2 GB 128K (full) 142–261
Qwen3-Coder-Next 79.7B, 3B active GGUF Q6_K 67.6 GB 256K (full) 171–321
Llama 3.1 70B 70.6B FP8 / INT8 75.5 GB 20K 15–21
Qwen3.6 35B-A3B 36.0B, 3B active FP16 / BF16 74.3 GB 256K (full) 86–153
Nemotron 3 Nano 30B-A3B 31.6B, 3.5B active FP16 / BF16 65.3 GB 256K (full) 77–135
Gemma 4 31B 31.3B FP16 / BF16 65.8 GB 256K (full) 17–24
GLM-4.7 Flash 31.2B, 3B active FP16 / BF16 64.9 GB 198K (full) 83–147
Xing 4.0 29B-A4B 31.2B, 4B active FP16 / BF16 64.8 GB 256K (full) 66–115
Qwen3 30B-A3B 30.5B, 3.3B active FP16 / BF16 63.9 GB 40K (full) 73–129
Muse Glimmer 30B 29.8B FP16 / BF16 61.7 GB 128K (full) 18–25
Qwen3.6 27B 27.8B FP16 / BF16 58.0 GB 256K (full) 19–27
Qwen3.8 27B 27.8B FP16 / BF16 58.0 GB 256K (full) 19–27
Gemma 4 26B-A4B 25.8B, 4B active FP16 / BF16 53.7 GB 256K (full) 66–116
gpt-oss-20b 20.9B, 3.6B active FP16 / BF16 43.6 GB 128K (full) 74–129
Gemma 4 12B 12.0B FP16 / BF16 25.4 GB 256K (full) 43–61
Qwen3.5 9B 9.7B FP16 / BF16 20.6 GB 256K (full) 53–75
Qwen3 8B 8.2B FP16 / BF16 18.5 GB 40K (full) 58–83
Llama 3.1 8B 8.0B FP16 / BF16 18.1 GB 128K (full) 60–85
Gemma 4 E4B 8.0B FP16 / BF16 17.0 GB 128K (full) 63–90
Nemotron 3 Nano 4B 4.0B FP16 / BF16 8.78 GB 256K (full) 115–173
MiniCPM5 2B 2.5B FP16 / BF16 6.02 GB 128K (full) 159–249

Too big for one card

How many A100 80GB cards these need at Q4_K_M in one tensor-parallel group; some need more than 8.

How these numbers are worked out

Memory is the weights at each precision plus the KV cache for 8,192 tokens in FP16 and runtime overhead (0.5 GB plus 10%), computed from each model's files on Hugging Face with the LLM VRAM Calculator. Speed is estimated from memory bandwidth and the parameters read per token with the LLM Speed Calculator. Both tools take any other model from Hugging Face.

Other GPUs and Macs

Model numbers read from Hugging Face on .