ToolSite

What LLMs can four RTX 3090 GPUs run?

Four RTX 3090 GPUs give a model 96 GB (4 × 24 GB) of memory and 3,744 GB/s of bandwidth in total. Of the 40 open models tracked here, 24 fit at Q4_K_M with an 8,192-token context, and 1 more at a lower precision. The speeds assume tensor parallelism, as in vLLM or SGLang; llama.cpp splits layers across cards by default and then writes at about the speed of one card.

Models that fit on 4× RTX 3090

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 Q3_K_M 90.8 GB 150K 159–336
Mistral Medium 3.5 128B 127.7B GGUF Q4_K_M 82.7 GB 39K 24–33
Qwen3.5 122B-A10B 125.1B, 10B active GGUF Q5_K_M 91.5 GB 123K 100–193
Nemotron 3 Super 120B-A12B 123.6B, 12B active GGUF Q5_K_M 90.3 GB 256K (full) 90–170
gpt-oss-120b 116.8B, 5.1B active GGUF Q5_K_M 85.6 GB 128K (full) 143–295
Qwen3-Coder-Next 79.7B, 3B active GGUF Q8_0 87.4 GB 256K (full) 153–322
Llama 3.1 70B 70.6B GGUF Q8_0 80.0 GB 49K 24–35
Qwen3.6 35B-A3B 36.0B, 3B active FP16 / BF16 74.3 GB 256K (full) 111–218
Nemotron 3 Nano 30B-A3B 31.6B, 3.5B active FP16 / BF16 65.3 GB 256K (full) 102–197
Gemma 4 31B 31.3B FP16 / BF16 65.8 GB 256K (full) 29–42
GLM-4.7 Flash 31.2B, 3B active FP16 / BF16 64.9 GB 198K (full) 108–211
Xing 4.0 29B-A4B 31.2B, 4B active FP16 / BF16 64.8 GB 256K (full) 91–173
Qwen3 30B-A3B 30.5B, 3.3B active FP16 / BF16 63.9 GB 40K (full) 99–190
Muse Glimmer 30B 29.8B FP16 / BF16 61.7 GB 128K (full) 31–44
Qwen3.6 27B 27.8B FP16 / BF16 58.0 GB 256K (full) 33–47
Qwen3.8 27B 27.8B FP16 / BF16 58.0 GB 256K (full) 33–47
Gemma 4 26B-A4B 25.8B, 4B active FP16 / BF16 53.7 GB 256K (full) 92–175
gpt-oss-20b 20.9B, 3.6B active FP16 / BF16 43.6 GB 128K (full) 99–190
Gemma 4 12B 12.0B FP16 / BF16 25.4 GB 256K (full) 65–100
Qwen3.5 9B 9.7B FP16 / BF16 20.6 GB 256K (full) 77–121
Qwen3 8B 8.2B FP16 / BF16 18.5 GB 40K (full) 83–132
Llama 3.1 8B 8.0B FP16 / BF16 18.1 GB 128K (full) 85–135
Gemma 4 E4B 8.0B FP16 / BF16 17.0 GB 128K (full) 88–142
Nemotron 3 Nano 4B 4.0B FP16 / BF16 8.78 GB 256K (full) 135–239
MiniCPM5 2B 2.5B FP16 / BF16 6.02 GB 128K (full) 164–311

Too big for 4× RTX 3090

How many RTX 3090 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 .