# How much VRAM do you need to run an LLM? | What Can I Run AI > VRAM needed for 1B, 3B, 4B, 8B, 14B, 20B, 27B, 32B, 49B, 70B, 120B, 235B, 400B, 671B and 1T LLMs at Q4, Q8 and FP16, plus the cheapest GPU and Mac that fit each size. Source: https://whatcanirunai.vercel.app/how-much-vram/ # How much VRAM do you need to run an LLM? It comes down to parameter count and quantization. Pick a size for the full breakdown: every quant, long-context needs, and the cheapest hardware that fits. __Memory needed by model size and cheapest hardware at Q4__ | Size | Q4\_K\_M | Q8\_0 | FP16 | Cheapest GPU at Q4 | Cheapest Mac / mini PC at Q4 | | ----------------------------------------------------------------------------- | -------- | --------- | --------- | ----------------------------------------------------------------------------- | ------------------------------------------------------------------------------------- | | [1B–2B models](https://whatcanirunai.vercel.app/how-much-vram/1b/)e.g. Llama-3.2-1B-Instruct | 1.9 GB | 2.5 GB | 3.8 GB | [Intel Arc B570](https://whatcanirunai.vercel.app/hardware/arc-b570/)$219 | [Mac mini M4 16GB](https://whatcanirunai.vercel.app/hardware/mac-mini-m4-16/)$789 | | [2B–3B models](https://whatcanirunai.vercel.app/how-much-vram/3b/)e.g. Llama-3.2-3B-Instruct | 3.2 GB | 4.9 GB | 8.4 GB | [Intel Arc B570](https://whatcanirunai.vercel.app/hardware/arc-b570/)$219 | [Mac mini M4 16GB](https://whatcanirunai.vercel.app/hardware/mac-mini-m4-16/)$789 | | [4B–6B models](https://whatcanirunai.vercel.app/how-much-vram/4b/)e.g. Qwen3-4B | 3.8 GB | 5.9 GB | 10.3 GB | [Intel Arc B570](https://whatcanirunai.vercel.app/hardware/arc-b570/)$219 | [Mac mini M4 16GB](https://whatcanirunai.vercel.app/hardware/mac-mini-m4-16/)$789 | | [7B–8B models](https://whatcanirunai.vercel.app/how-much-vram/8b/)e.g. Qwen3-8B | 6.7 GB | 11.0 GB | 19.8 GB | [Intel Arc B570](https://whatcanirunai.vercel.app/hardware/arc-b570/)$219 | [Mac mini M4 16GB](https://whatcanirunai.vercel.app/hardware/mac-mini-m4-16/)$789 | | [12B–14B models](https://whatcanirunai.vercel.app/how-much-vram/14b/)e.g. Qwen3-14B | 11.3 GB | 19.0 GB | 35.0 GB | [Radeon RX 7600 XT](https://whatcanirunai.vercel.app/hardware/rx-7600-xt/)$329 | [Mac mini M6 24GB](https://whatcanirunai.vercel.app/hardware/mac-mini-m6-24/)$1,299 | | [16B–21B models](https://whatcanirunai.vercel.app/how-much-vram/20b/)e.g. gpt-oss-20b | 15.6 GB | 26.6 GB | 49.1 GB | [Radeon RX 7900 XT](https://whatcanirunai.vercel.app/hardware/rx-7900-xt/)$899 | [Mac mini M6 24GB](https://whatcanirunai.vercel.app/hardware/mac-mini-m6-24/)$1,299 | | [24B–27B models](https://whatcanirunai.vercel.app/how-much-vram/27b/)e.g. Qwen3.8-27B | 20.4 GB | 34.9 GB | 64.9 GB | [Intel Arc Pro B70 32GB](https://whatcanirunai.vercel.app/hardware/arc-pro-b70/)$949 | [Mac mini M6 32GB](https://whatcanirunai.vercel.app/hardware/mac-mini-m6-32/)$1,799 | | [30B–32B models](https://whatcanirunai.vercel.app/how-much-vram/32b/)e.g. Qwen3-32B | 23.8 GB | 41.0 GB | 76.4 GB | [Intel Arc Pro B70 32GB](https://whatcanirunai.vercel.app/hardware/arc-pro-b70/)$949 | [Mac mini M5 Pro 64GB](https://whatcanirunai.vercel.app/hardware/mac-mini-m5-pro-64/)$3,199 | | [40B–50B models](https://whatcanirunai.vercel.app/how-much-vram/49b/)e.g. Llama-3\_3-Nemotron-Super-49B-v1\_5 | 35.8 GB | 61.9 GB | 115.7 GB | [NVIDIA RTX A6000](https://whatcanirunai.vercel.app/hardware/rtx-a6000/)$4,649 | [Mac mini M5 Pro 64GB](https://whatcanirunai.vercel.app/hardware/mac-mini-m5-pro-64/)$3,199 | | [70B models](https://whatcanirunai.vercel.app/how-much-vram/70b/)e.g. Llama-3.3-70B-Instruct | 50.2 GB | 87.2 GB | 163.3 GB | [NVIDIA RTX Pro 6000 Blackwell 96GB](https://whatcanirunai.vercel.app/hardware/rtx-pro-6000-blackwell/)$8,565 | [Framework Desktop (Ryzen AI Max+ 395) 128GB](https://whatcanirunai.vercel.app/hardware/framework-desktop-128/)$3,449 | | [120B models](https://whatcanirunai.vercel.app/how-much-vram/120b/)e.g. gpt-oss-120b | 82.5 GB | 143.8 GB | 269.7 GB | [NVIDIA RTX Pro 6000 Blackwell 96GB](https://whatcanirunai.vercel.app/hardware/rtx-pro-6000-blackwell/)$8,565 | [Framework Desktop (Ryzen AI Max+ 395) 128GB](https://whatcanirunai.vercel.app/hardware/framework-desktop-128/)$3,449 | | [200B–300B models](https://whatcanirunai.vercel.app/how-much-vram/235b/)e.g. Qwen3-235B-A22B | 164.9 GB | 288.3 GB | 541.7 GB | Needs 2+ GPUs | [Mac Studio M5 Ultra 256GB](https://whatcanirunai.vercel.app/hardware/mac-studio-m5-ultra-256/)$9,499 | | [300B–480B models](https://whatcanirunai.vercel.app/how-much-vram/400b/)e.g. Qwen3.5-397B-A17B | 282.2 GB | 493.9 GB | 928.8 GB | Needs 2+ GPUs | [Mac Studio M5 Ultra 512GB](https://whatcanirunai.vercel.app/hardware/mac-studio-m5-ultra-512/)$22,299 | | [500B–800B models](https://whatcanirunai.vercel.app/how-much-vram/671b/)e.g. DeepSeek-R1 | 478.2 GB | 837.4 GB | 1575.3 GB | Needs 2+ GPUs | None we track | | [1T+ models](https://whatcanirunai.vercel.app/how-much-vram/1t/)e.g. Kimi-K2-Instruct | 716.6 GB | 1255.1 GB | 2361.7 GB | Needs 2+ GPUs | None we track | Estimates at a \~4k context from [our formula](https://whatcanirunai.vercel.app/method/). Prices are dated snapshots (GPUs at list price). Exact model? Look it up in [all models](https://whatcanirunai.vercel.app/models/) or use the [VRAM calculator](https://whatcanirunai.vercel.app/vram-calculator/). ## Quick answers How much VRAM do I need to run an LLM? At Q4\_K\_M, plan on roughly 0.7 GB per billion parameters plus about 1 GB at a \~4k context: 1B–2B \~1.9 GB, 2B–3B \~3.2 GB, 4B–6B \~3.8 GB, 7B–8B \~6.7 GB, 12B–14B \~11.3 GB, 16B–21B \~15.6 GB, 24B–27B \~20.4 GB, 30B–32B \~23.8 GB, 40B–50B \~35.8 GB, 70B \~50.2 GB, 120B \~82.5 GB, 200B–300B \~164.9 GB, 300B–480B \~282.2 GB, 500B–800B \~478.2 GB, 1T+ \~716.6 GB. Q8 needs about 1.75× that and FP16 about 3.3×. Is 16 GB of VRAM enough for local LLMs? Yes for models up to about 14B at Q4\_K\_M: a 14B model needs about 11.3 GB at a \~4k context, which leaves room for longer context. gpt-oss-20b needs about 15.6 GB at Q4\_K\_M by our formula, so it is tight on a 16 GB card, with little room for context; a 20 GB card runs it comfortably. 24B–27B dense models need about 20.4 GB at Q4\_K\_M, so 24 GB or more. Is 24 GB of VRAM enough for a 70B model? Not at Q4\. A 70B model needs about 50.2 GB at Q4\_K\_M. You need two 24 GB cards, a 128 GB unified-memory machine, or a much smaller quant. [Ollama VRAM requirements by tag](https://whatcanirunai.vercel.app/ollama-vram-requirements/) · [GGUF VRAM by quant (llama.cpp, LM Studio)](https://whatcanirunai.vercel.app/gguf-vram-requirements/) · [Buying guides](https://whatcanirunai.vercel.app/best-gpu-for/) · [Best LLMs by VRAM size](https://whatcanirunai.vercel.app/best-llm-for/) · [Best GPU for local LLMs](https://whatcanirunai.vercel.app/best-gpu-for-local-llm/) --- Agent guide: https://whatcanirunai.vercel.app/llms.txt · JSON API: https://whatcanirunai.vercel.app/developers/