Source: https://getrigbase.com/guides/how-much-system-ram-for-local-llms/ Last updated: 2026-10-28 # How much system RAM do you need for local LLMs? Quick answer **How much system RAM do you need for local LLMs?** Less than most guides claim — if the model weights and KV cache fit in GPU (or unified) memory. Decode then streams from VRAM/HBM bandwidth, not from DDR. 32 GB of system RAM is enough for a multi-GPU rig that keeps the whole model on the cards. You need more RAM when you offload layers/experts to system memory or run big contexts with CPU involvement. Last updated: 28 Oct 2026 Most “you need 128 GB of RAM” advice assumes CPU offload. When the model fits in VRAM, system RAM matters much less than GPU memory and bandwidth. **Quick answer:** If the model is fully on the GPU(s), **32 GB of system RAM is often enough** — including multi-GPU 3090 rigs. Spend the budget on VRAM (or a used 3090 stack) first. Add RAM when you offload experts to system memory or want huge CPU-side context. ## Why VRAM beats system RAM for decode speed Generation speed is limited by how fast the runtime can read weights for each token. GPU memory bandwidth is hundreds of GB/s to TB/s; DDR5 is far lower. So a model that spills to system RAM usually drops to a fraction of the all-GPU speed. The [VRAM calculator](https://getrigbase.com/can-it-run/) separates weight size from KV-cache size so you can see what must fit on the card. ## 32 GB system RAM is a working multi-GPU setup An 8× RTX 3090 rig (192 GB VRAM) with only **32 GB DDR5 ECC RDIMM** works well for models that stay entirely on the GPUs — the host RAM just needs to hold the OS, the runtime, and staging buffers. That is the opposite of “buy 256 GB first” advice, and it is a differentiator vs cloud-shaped shopping lists. Example from RigBase: [EPYC 7F52 jank rig: 4x3090 + 4090 + 4080 (136 GB VRAM)](https://getrigbase.com/builds/epyc-7f52-6gpu-136gb/) lists **32 GB RDIMM 3200** with 4x RTX 3090, RTX 4090, RTX 4080 running Llama 3.1 70B at 9 tok/s (Q8_0). ## Measured builds that state ~32 GB of system RAM | GPUs | System RAM | Model | Quant | tok/s | |---|---|---|---|---| | [Apple M6 (12-core GPU, unified)](https://getrigbase.com/builds/mac-mini-m6-32gb-lm-studio-oct-2026/) | 32 GB unified | [Qwen3.6 35B-A3B](https://getrigbase.com/models/qwen3-6-35b-a3b/) | MLX 4-bit | **63.2** | | [RTX 3090 24GB](https://getrigbase.com/builds/rtx-3090-250w-windows-qwen38-27b-mtp-club3090/) | 32 GB DDR5 | [Qwen3.8 27B](https://getrigbase.com/models/qwen3-8-27b/) | Unsloth UD-Q4_K_M (nextn/MTP head em | **56.37** | | [RX 9070 XT 16 GB + Radeon AI PRO R9700 32 GB](https://getrigbase.com/builds/9800x3d-rx-9070-xt-plus-r9700-rocm-llamacpp/) | 32 GB DDR5-6000 | [Qwen3.8 27B](https://getrigbase.com/models/qwen3-8-27b/) | IQ4_XS-v3, q8_0 KV | **31.49** | | [RTX 5070 12GB](https://getrigbase.com/builds/rtx-5070-12gb-qwen38-flash-next-expert-streaming/) | 32 GB DDR4-2400 | [Qwen3.8 Flash-Next](https://getrigbase.com/models/qwen3-8-flash-next/) | UD-IQ3_XXS, experts streamed from RA | **11.5** | | [4x RTX 3090, RTX 4090, RTX 4080](https://getrigbase.com/builds/epyc-7f52-6gpu-136gb/) | 32 GB RDIMM 3200 | [Llama 3.1 70B](https://getrigbase.com/models/llama-3-1-70b/) legacy | Q8_0 | **9** | ## When you do need lots of RAM - **Expert offload:** Qwen3.8 Flash-Next on an RTX 5070 12GB streamed experts from RAM/SSD at 11.5 tok/s (UD-IQ3_XXS) — usable, but not 100+ tok/s. [Source build](https://getrigbase.com/builds/rtx-5070-12gb-qwen38-flash-next-expert-streaming/). - **Big MoE on few cards:** the 6×3090 Flash-Next build uses 192 GB DDR5 alongside 144 GB VRAM. - **Unified memory boxes:** Mac Studio / Strix Halo / DGX Spark share one pool — size that pool for weights + KV, not “VRAM + RAM” separately. Next: [multi-GPU 3090 guide](https://getrigbase.com/guides/multi-gpu-rtx-3090-local-llm/) · [Is Q4 good enough?](https://getrigbase.com/guides/quantization-is-q4-good-enough/) · [best LLMs by VRAM](https://getrigbase.com/best/) ## FAQ ### How much system RAM do you need for local LLMs? Less than most guides claim — if the model weights and KV cache fit in GPU (or unified) memory. Decode then streams from VRAM/HBM bandwidth, not from DDR. 32 GB of system RAM is enough for a multi-GPU rig that keeps the whole model on the cards. You need more RAM when you offload layers/experts to system memory or run big contexts with CPU involvement. ### Can an 8× RTX 3090 run with only 32 GB of RAM? Yes, when the model is fully resident on the GPUs. Builders run 8×3090 (192 GB VRAM) with 32 GB DDR5 ECC RDIMM successfully for that case. If you stream experts from system RAM (common for 100B+ MoE on fewer cards), plan 128–256 GB instead — see the 6×3090 Flash-Next build with 192 GB DDR5. ### When does more RAM help? Expert/layer offload, very long context with CPU KV, running multiple models, or OS + browser + IDE alongside. Unified-memory boxes (Mac, Strix Halo, DGX Spark) share one pool, so “system RAM” and “VRAM” are the same budget. Ad · this space is available **Sponsor this spot: reach people building local AI rigs.** One clearly-labeled slot per page, no tracking scripts. [Sponsor this spot →](https://getrigbase.com/advertise/#inquiry) Diese Seite auf Deutsch: [Wie viel System-RAM für lokale LLMs?](https://getrigbase.com/de/guides/wie-viel-system-ram-fuer-lokale-llms/)