ToolSite

What LLMs can an H200 run?

An H200 gives a model 141 GB of memory and 4,800 GB/s of bandwidth. Of the 40 open models tracked here, 25 fit at Q4_K_M with an 8,192-token context, and 5 more at a lower precision.

Models that fit on one H200

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
GLM-5.3 Flash 321.3B, 18B active GGUF Q2_K 138 GB 199K 147–271
MiMo V2.6 Flash 310.8B, 15B active GGUF Q2_K 134 GB 296K 166–312
DeepSeek V4 Flash 0731 304.2B, 13B active GGUF Q2_K 132 GB 108K 173–326
DeepSeek V4 Flash 290.9B, 13B active GGUF Q2_K 126 GB 169K 173–326
MiniMax M2.7 228.7B, 10B active GGUF Q3_K_M 117 GB 97K 158–294
Qwen3.8 Flash Next 180.0B, 6B active GGUF Q5_K_M 131 GB 256K (full) 218–424
Mistral Medium 3.5 128B 127.7B As published (FP8) 140 GB 9K 19–26
Qwen3.5 122B-A10B 125.1B, 10B active GGUF Q8_0 137 GB 168K 111–200
Nemotron 3 Super 120B-A12B 123.6B, 12B active GGUF Q8_0 135 GB 256K (full) 96–171
gpt-oss-120b 116.8B, 5.1B active GGUF Q8_0 128 GB 128K (full) 183–347
Qwen3-Coder-Next 79.7B, 3B active GGUF Q8_0 87.4 GB 256K (full) 260–523
Llama 3.1 70B 70.6B GGUF Q8_0 80.0 GB 128K (full) 32–45
Qwen3.6 35B-A3B 36.0B, 3B active FP16 / BF16 74.3 GB 256K (full) 173–326
Nemotron 3 Nano 30B-A3B 31.6B, 3.5B active FP16 / BF16 65.3 GB 256K (full) 156–291
Gemma 4 31B 31.3B FP16 / BF16 65.8 GB 256K (full) 39–55
GLM-4.7 Flash 31.2B, 3B active FP16 / BF16 64.9 GB 198K (full) 167–314
Xing 4.0 29B-A4B 31.2B, 4B active FP16 / BF16 64.8 GB 256K (full) 137–251
Qwen3 30B-A3B 30.5B, 3.3B active FP16 / BF16 63.9 GB 40K (full) 151–279
Muse Glimmer 30B 29.8B FP16 / BF16 61.7 GB 128K (full) 41–58
Qwen3.6 27B 27.8B FP16 / BF16 58.0 GB 256K (full) 44–62
Qwen3.8 27B 27.8B FP16 / BF16 58.0 GB 256K (full) 44–62
Gemma 4 26B-A4B 25.8B, 4B active FP16 / BF16 53.7 GB 256K (full) 138–253
gpt-oss-20b 20.9B, 3.6B active FP16 / BF16 43.6 GB 128K (full) 151–279
Gemma 4 12B 12.0B FP16 / BF16 25.4 GB 256K (full) 93–138
Qwen3.5 9B 9.7B FP16 / BF16 20.6 GB 256K (full) 112–168
Qwen3 8B 8.2B FP16 / BF16 18.5 GB 40K (full) 123–186
Llama 3.1 8B 8.0B FP16 / BF16 18.1 GB 128K (full) 125–190
Gemma 4 E4B 8.0B FP16 / BF16 17.0 GB 128K (full) 131–201
Nemotron 3 Nano 4B 4.0B FP16 / BF16 8.78 GB 256K (full) 219–364
MiniCPM5 2B 2.5B FP16 / BF16 6.02 GB 128K (full) 282–501

Too big for one card

How many H200 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 .