The right Personal Computer for running AI at home
Check whether a model fits your hardware, find the cheapest machine that runs it, and read guides where every spec and price is cited.
Can My PC Run It?
Pick a model, quantization and context length. See the memory it needs, whether it fits your GPU or unified memory, and a speed ceiling.
AI PC Builder
Set a budget and target model. Get every GPU, dual-GPU or compact system that fits, sorted by cited price, with a parts outline.
Guides
Memory, quantization and the hardware basics behind local AI. View all →
What Is VRAM and How Much Do You Need? (AI and Gaming)
VRAM explained: what video memory does, why it limits AI models and games, and how much you need for local LLMs…
Read →GuidesWhat Is VRAM and How Much Do You Need for Local LLMs?
VRAM needed to run LLMs locally by model size and quantization, from 4B to 671B, with the cheapest GPU or system…
Read →GuidesVRAM vs RAM
VRAM vs RAM explained: where each sits, how fast each is, and why VRAM matters more for local AI and gaming.…
Read →GuidesHow Much RAM Do I Need for a PC in 2026?
How much RAM for a PC: practical guidance for everyday use, gaming, creative work and running AI models…
Read →GuidesHow Much RAM for Local AI and LLMs (Is 128GB RAM Worth It?)
How much system RAM you need to run LLMs locally, and when 128GB RAM makes sense: with a GPU, on CPU only, or on…
Read →GuidesLLM Quantization Explained
What LLM quantization is, how Q4, Q5, Q8 and FP16 change VRAM use, and how to pick a level for your…
Read →GuidesUnified Memory vs VRAM for LLMs
How unified memory compares to discrete GPU VRAM for running LLMs: capacity, bandwidth and what fits on Apple…
Read →GuidesWhat Is an AI PC? NPUs Explained (Neutral Guide)
What an AI PC is, what an NPU does, how Copilot+ PCs fit in, and what matters if you want to run AI models…
Read →GuidesWhat Is a Copilot+ PC? (Neutral Explainer)
What Microsoft means by a Copilot+ PC, the hardware requirements, and how it relates to running AI models…
Read →GuidesOCuLink and eGPUs
OCuLink explained, OCuLink vs Thunderbolt for an external GPU for a laptop or mini PC, and what to check in an…
Read →GuidesMini PC and SFF Cases
How to choose a mini PC or small form factor (SFF) case: motherboard size, GPU clearance, cooling and power supply…
Read →GuidesHow Much Does a Personal Computer Cost in 2026?
Personal computer prices in 2026 by use: everyday PCs, gaming PCs, and AI-capable desktops and mini PCs, with…
Read →GuidesRAM Prices 2026
RAM price tracker: why DDR5 RAM prices rose in 2026, what memory makers and analysts forecast, and whether RAM…
Read →Can I run it?
Memory needed by model family, with the cheapest hardware that fits. View all →
Can I Run DeepSeek Locally? Hardware Requirements
Hardware needed to run DeepSeek-R1 and its distilled models locally, from 8B to the full 671B, with memory…
Read →Can I run it?How to Run Qwen3 Locally
Memory and hardware needed to run Qwen3 models locally, from 4B to 235B-A22B, including the mixture-of-experts…
Read →Can I run it?How Much VRAM for Llama? 8B, 70B and 405B
VRAM needed to run Llama 3.1 and 3.3 locally at different quantizations, with the cheapest GPU or system that…
Read →Can I run it?gpt-oss Hardware Requirements
What you need to run OpenAI gpt-oss-20b and gpt-oss-120b locally, using OpenAI's published memory…
Read →Can I run it?Gemma 3 VRAM Requirements
VRAM needed to run Google Gemma 3 models locally at Q4 and Q8, and which GPUs…
Read →Can I run it?Running 70B Models at Home
What it takes to run a 70B-parameter LLM at home: dual GPUs, 128 GB unified-memory mini PCs or a Mac…
Read →GPUs
Graphics cards for local LLMs, new and used. View all →
Best GPU for Local LLMs (by Budget)
How to choose a GPU for running LLMs locally: VRAM first, then bandwidth, with cited specs and launch prices for…
Read →GPUsBest Used GPUs for AI
Why used 24 GB GPUs like the RTX 3090 are popular for local LLMs, what to check before buying, and how they…
Read →GPUsGeForce RTX 5060 VRAM
The GeForce RTX 5060 has 8 GB of VRAM. What that means for games and local AI models, with cited…
Read →GPUsGeForce RTX 5070 VRAM
The GeForce RTX 5070 has 12 GB of VRAM. What that means for games and local AI models, with cited…
Read →GPUsGeForce RTX 5080 VRAM
The GeForce RTX 5080 has 16 GB of VRAM. What that means for games and local AI models, with cited…
Read →GPUsBest Laptops for Local LLMs
How to pick a laptop for running LLMs locally: GPU VRAM, unified memory, and the trade-offs of mobile GPUs and…
Read →Builds
Desktop builds by budget, with cited GPU prices. View all →
$1,000 AI PC Build for Local LLMs
A budget AI PC build around $1,000 for running local LLMs, built around a 16 GB GPU, with cited GPU…
Read →Builds$2,000 AI PC Build for LLMs
A $2,000-class PC build for running LLMs locally, built around a 16 GB RTX 5070 Ti or…
Read →Builds$5,000 AI Workstation Build
A $5,000-class AI workstation build for local LLMs around the 32 GB RTX…
Read →BuildsDual-GPU LLM Build
How to build a dual-GPU PC for local LLMs: two 24 GB cards for 70B models, and what to…
Read →Compare
Head-to-head comparisons of the main options. View all →
NVIDIA DGX Spark vs AMD Strix Halo (Ryzen AI Max+ 395) vs Mac Studio
NVIDIA DGX Spark vs AMD Strix Halo (Ryzen AI Max+ 395) vs Apple Mac Studio (M5 Max) compared for running LLMs:…
Read →CompareMac mini vs Mini PC for Local AI
Mac mini (M6 and M5 Pro) compared with 128 GB Ryzen AI Max mini PCs for running LLMs…
Read →CompareRTX 5090 vs 4090 vs 3090 for LLMs
RTX 5090, 4090 and 3090 compared for local LLMs on VRAM, memory bandwidth and launch…
Read →CompareOllama vs LM Studio vs llama.cpp
Ollama, LM Studio and llama.cpp compared from a hardware perspective: platforms, GPU backends and what affects…
Read →Mini PCs
Compact systems with large unified memory. View all →
AI Mini PC Guide
AI mini PC buying guide: the best mini PC for AI compared on memory, bandwidth and cited price — Ryzen AI Max+ 395…
Read →Mini PCsRyzen AI Max Mini PCs Compared
AMD Ryzen AI Max (AI Max+ 395, Strix Halo) mini PCs compared: 128GB and 64GB configurations, cited prices and…
Read →Mini PCsStrix Halo (AMD Ryzen AI Max) Mini PCs Compared
AMD Strix Halo explained: Ryzen AI Max / AI Max+ 395 specs and the 128GB and 64GB mini PCs that use it, compared…
Read →Mini PCsGMKtec EVO-X2
GMKtec EVO-X2 Ryzen AI Max+ 395 mini PC: cited specs, reported price, and what models fit in 128…
Read →Mini PCsMinisforum MS-S1 Max
Minisforum MS-S1 Max Ryzen AI Max+ 395 mini PC: cited specs, reported price, and what models…
Read →Mini PCsBeelink GTR9 Pro
Beelink GTR9 Pro Ryzen AI Max+ 395 mini PC: cited specs, reported price, and what models…
Read →Mini PCsFramework Desktop for AI
Framework Desktop with Ryzen AI Max+ 395 for local LLMs: cited specs, reported price and stock, and what…
Read →Mini PCsCorsair AI Workstation 300
Corsair AI Workstation 300: Ryzen AI Max+ 395, 128GB unified memory, cited price and what…
Read →Mini PCsDGX Spark Price and Partner Systems
NVIDIA DGX Spark price as reported, specs, and partner GB10 systems, with sources and…
Read →Benchmarks
Methodology and theoretical ceilings until we test. View all →
Popular models at a glance
Estimated memory at 8k context and the cheapest single device in our data that fits.
| Model | Params (total / active) | Est. memory, Q4 @ 8k ctx | Est. memory, Q8 @ 8k ctx | Cheapest fit in our data (Q4) |
|---|---|---|---|---|
| Llama 3.1 8B | 8.03B / 8.03B | 7.0 GB | 10.6 GB | GeForce RTX 5060 ($299, GPU launch MSRP) |
| Qwen3 30B-A3B (MoE) | 30.5B / 3.3B | 21.1 GB | 34.9 GB | GeForce RTX 3090 ($1,499, GPU launch MSRP) |
| gpt-oss-20b (MoE) | 21B / 3.6B | 13.7 GB | 24.8 GB | GeForce RTX 5060 Ti 16GB ($429, GPU launch MSRP) |
| Gemma 3 27B | 27.4B / 27.4B | 19.2 GB | 31.6 GB | GeForce RTX 3090 ($1,499, GPU launch MSRP) |
| Llama 3.3 70B | 70.6B / 70.6B | 46.8 GB | 78.7 GB | GMKtec EVO-X2 (128GB / 1TB) ($3,499.99, system, listed price) |
| gpt-oss-120b (MoE) | 117B / 5.1B | 64.7 GB | 126.8 GB | GMKtec EVO-X2 (128GB / 1TB) ($3,499.99, system, listed price) |
Parameter counts: Google: Gemma 3 model card; Hugging Face: openai/gpt-oss-120b model card; Hugging Face: meta-llama/Llama-3.1-8B; Hugging Face: meta-llama/Llama-3.3-70B-Instruct; Hugging Face: Qwen3 model cards (e.g. Qwen/Qwen3-30B-A3B). Memory figures are our estimates (see methodology), not measurements.