Best Local AI Hardware: Builds from $600 to $3,499
Updated: November 2, 2026
Five complete build tiers from a CPU-only starter to a 32GB RTX 5090 workstation: the exact part lists, the memory-bandwidth math that decides how fast each one generates tokens, and which tier gives the best value for your budget.
Need software next? Explore the models directory for downloads, grab optimized picks from the 8GB model guide, and keep our Ollama troubleshooting guide handy while you build. Pricing a machine? The best PC for local LLMs guide has three parts lists priced November 2026, and the best local AI server guide covers headless and plug-and-play boxes. Ready to master AI? Follow our AI Learning Path to learn what AI really is, not just how to use it.
Budget Builds
$600-$900 at MSRP • Runs 34 directory models (up to 7B)
Performance Builds
$1,200-$2,500 at MSRP • Runs 53 directory models (up to 14B on 16GB; 27-34B needs a 24GB card)
Enterprise Builds
$5,000 and up • Check model and multi-GPU requirements
Bought the hardware? Now get the most out of it.
Every course on running local models, RAG, agents and fine-tuning, so the machine you just speced actually earns its price.
Want to go deeper than this article?
Free account unlocks the first chapter of all 25 courses — RAG, agents, MCP, voice AI, MLOps, real GitHub repos.
Cheapest AI PC: The Quick Answer
The cheapest unified-memory AI PC is the base Mac mini with the M6 chip and 16GB of unified memory (from $899 on Apple's store on November 2, 2026): it runs 7-8B models well out of the box with zero build effort. The cheapest discrete-GPU route is a used RTX 3060 12GB build, which handles everything up to the 14B class and leaves you an upgrade path.
One buying note: Apple moved the Mac mini line to the M6 and M5 Pro chips in September 2026, so older guides still quoting M4 configurations and prices are out of date. Apple lists 153GB/s of memory bandwidth for the entry M6 configuration, 170GB/s for the 24GB and 32GB M6 configurations, and 307GB/s for the M5 Pro. Check the configuration you are buying against Apple's official Mac mini tech specs before you order, since memory is not upgradeable after purchase. For the Mac route, see our Apple Silicon AI guide; for the RTX 3060 route, the budget local AI machine guide has the full parts list, and this page shows exactly what a 12GB card runs. Already own a card? We keep per-GPU model picks (RTX 3060 to 5090, RX 7900 XTX) and per-Mac picks, and a full local AI server comparison by budget.
Local AI PC: What to Buy
A local AI PC is an ordinary desktop with one job that other PCs do not have: holding an entire model in graphics memory. If the model's weights do not fit in VRAM, the card sits half idle while the CPU and system RAM do the work several times slower. So pick the graphics card first, by memory size, and spend what is left on everything else.
- →16GB of VRAM: RTX 5060 Ti 16GB ($429 MSRP) or, if you want more speed, RTX 5070 Ti 16GB ($749 MSRP). AMD's RX 9070 XT 16GB ($599 SEP) sits between them on bandwidth; confirm it in Ollama's GPU list before buying. Runs the whole 7B-14B class at Q4 with room for context, plus 20B mixture-of-experts models such as gpt-oss:20b.
- →24GB of VRAM: a used RTX 3090 24GB. No current GeForce card offers 24GB below the RTX 5090's 32GB, so the used 3090 is still the way into the 27B-32B class, including Qwen3.6-27B at Q4.
- →Everything else: a current 6-core CPU (Ryzen 5 9600X class), 32GB of DDR5, a 1TB NVMe drive and a power supply sized for the card. None of these change how fast a GPU-resident model generates tokens.
Best PC for Local AI: 16GB or 24GB?
The best PC for local AI is the one whose VRAM matches the largest model you actually want to run, not the one with the fastest CPU. Generating a token means reading every weight in the model once, so the card's memory size decides what fits and its memory bandwidth decides how fast; the CPU mostly waits. A 16-core CPU does nothing for a model that lives on the GPU, while doubling bandwidth doubles the speed ceiling.
- →Best at 16GB: RTX 5070 Ti 16GB (896GB/s) if speed matters, RTX 5060 Ti 16GB (448GB/s) if budget matters. Same models fit on both; the 5070 Ti has twice the ceiling.
- →Best at 24GB: used RTX 3090 (936GB/s). Unlocks the 27B-32B class, which is where the strongest single-GPU models live right now.
- →Best if money is no object: RTX 5090 32GB (1,792GB/s, $1,999 MSRP). Runs the 32B class with long context and roughly doubles the 3090's ceiling.
- →Why VRAM beats CPU: a Ryzen 5 9600X and a Ryzen 9 9950X give identical token speeds once the model sits on the card. Put the CPU money into the GPU's memory size, and 32GB of system RAM so large models can still load when they spill.
Best PC for Local LLM: The Concrete Pick
The best PC for running a local LLM, as of November 2, 2026, is a Ryzen 5 9600X with an RTX 5060 Ti 16GB, 32GB of DDR5-6000 and a 1TB NVMe drive: about $1,970 at that day's retail prices (the 16GB card was $799.99 on NVIDIA's marketplace, the memory kit $479.99 and the CPU $179.99 at Walmart). It runs the whole 8B-14B class with room for context and the 20B mixture-of-experts models. If the models you want are the 27B-32B class, the pick is the same tower with a used RTX 3090, the only 24GB card below an RTX 5090, which listed at $3,340 and up that day.
Every part, its price and the page it was read from is in the best PC for local LLMs guide (three builds: about $1,970, $2,840 and $5,800). If the machine will sit headless on the network instead, the best local AI server guide covers the plug-and-play boxes, the used-3090 server and the three-year cost against ChatGPT.
August 2026 Update: What the New Model Wave Means for Your Hardware
The summer 2026 wave of open models has kept splitting the hardware answer in two directions:
- →Small cards got better, not obsolete. Google's Gemma 4 family (Apache 2.0) runs offline on everything from a Raspberry Pi to a mid-range GPU — an 8-12GB card is still a genuinely useful AI machine.
- →24GB is the new sweet spot for one-GPU builds. Qwen3.6-27B — the current best single-GPU model — fits an RTX 4090/5090 at ~17GB. See what a 24GB card runs best.
- →No GPU at all? System RAM sets the ceiling instead — see the picks for 32GB of RAM and 64GB of RAM.
- →The new giants need server-class memory. GLM-5.2 (753B, MIT) is self-hostable — but quantized it wants a 256-512GB unified-memory Mac Studio or an EPYC-class rig, and the 2.8T Kimi K3 is out of home reach entirely. Don't buy a bigger GPU chasing them.
- →August did not move the hardware goalposts. Meta's Muse Glimmer and Alibaba's Qwen3.8-27B both land in the same 24GB class as the models above, and NVIDIA's Nemotron 3.5 Lightning is a 30B mixture-of-experts with only 3B active — so it runs far lighter than its parameter count suggests. None of them is a reason to buy a bigger card.
The builds below remain current — nothing in the new wave changes the price-to-capability picks. One caveat on timing: GPU street prices are running 50%+ over MSRP right now and the RTX 50 Super refresh is delayed — see why GPU prices are so high and what to buy right now.
5 AI PC Builds Compared
Five part lists from under $1,000 to about $3,500 at MSRP, the range most people actually shop in, sized against the models we recommend now: Qwen3.6-27B, the Qwen 14B class, and Gemma 4. Each card shows two numbers: the MSRP total, from NVIDIA's and AMD's launch prices, and what the same parts cost at retail on November 2, 2026. The gap is large, because GPU and DDR5 prices rose sharply in 2026; the retail figures are the prices read that day for the best PC for local LLMs guide, which lists every part and the page it was read from, and they will move, so check current prices on the day you buy. The speeds under each build are arithmetic ceilings: card memory bandwidth divided by the weight bytes a Q4_K_M model moves per token. Real output lands below those ceilings.
Budget Champion
under $1,000 at MSRPstreet, Nov 2 2026: ~$1,170
CPU: Ryzen 5 9600X (6-core)
RAM: 32GB DDR5-5600
GPU: None (CPU only)
Storage: 1TB NVMe
Street, Nov 2 2026: Ryzen 5 9600X $179.99, a 32GB DDR5-6000 kit $479.99 (the memory is now the biggest line), plus board, 1TB drive, PSU and case for ~$1,170.
📐 Bandwidth ceiling (Q4_K_M):
- • Llama 3.1 8B (~4.9GB): ~18 tok/s ceiling
- • Mistral 7B (~4.4GB): ~20 tok/s ceiling
- • Gemma 4 E2B/E4B: sized for this class
- Dual-channel DDR5-5600 ≈ 90GB/s; CPU-only decode lands furthest below its ceiling.
Verdict: Perfect starter. Handles all models in our 8GB guide smoothly.
Sweet Spot Build
about $1,100 to $1,300 at MSRPstreet, Nov 2 2026: ~$2,130
CPU: Ryzen 5 9600X (6-core)
RAM: 32GB DDR5-5600
GPU: RTX 5060 Ti 16GB ($429 MSRP; $799.99-$879.99 street, Nov 2 2026)
Storage: 2TB NVMe
Street total uses the budget parts list from the best PC for local LLMs guide (~$1,970 with a 1TB drive) with a 2TB drive ($339.99) in place of the 1TB.
📐 Bandwidth ceiling (Q4_K_M):
- • Llama 3.1 8B (~4.9GB): ~91 tok/s ceiling
- • Qwen 3 14B (~9.3GB): ~48 tok/s ceiling
- • Gemma 3 12B (~8.2GB): ~55 tok/s ceiling
- RTX 5060 Ti: 448GB/s, 180W (NVIDIA spec). 16GB leaves room for long context on the 14B class.
Verdict: Best bang-for-buck. The cheapest new 16GB card, so the whole 14B class fits with context to spare, and 20B MoE models such as gpt-oss:20b fit too. See what a 16GB card runs and full GPU comparisons. AMD alternative: the RX 9070 XT 16GB ($599 SEP) has more bandwidth at 640GB/s; confirm ROCm support in Ollama's GPU list first.
24GB on a Budget
about $1,500 with a used 3090 at earlier pricesstreet, Nov 2 2026: ~$3,050-$3,320
CPU: Ryzen 7 9700X (8-core)
RAM: 64GB DDR5-5600
GPU: RTX 3090 24GB (used, previous generation; $1,299-$1,568 used, Nov 1-2 2026)
Storage: 1TB NVMe
Street, Nov 2 2026: ~$1,755 for the platform (9700X $334.99, 64GB DDR5-6000 kit $833.63, board, 1TB drive, 850W PSU, case) plus the card. The used band is an eBay listing at $1,299 and resaleprices.com's $1,568 average (1 Nov 2026); price yours from sold listings, not asking prices.
📐 Bandwidth ceiling (Q4_K_M):
- • Qwen3.6-27B (~17.4GB): ~54 tok/s ceiling
- • Qwen 2.5 Coder 32B (~19.9GB): ~47 tok/s ceiling
- • Card power: 350W board TGP (NVIDIA spec)
- RTX 3090: 936GB/s — the cheapest high-bandwidth 24GB card.
Verdict: The one previous-generation card on this page, kept on purpose: no current GeForce card offers 24GB below the RTX 5090's 32GB (the RTX 5080 is 16GB), so the used 3090 is still the cheapest route to 24GB, and 24GB runs the best single-GPU models at full Q4. See what 24GB unlocks. Used prices move week to week, so check current sold listings before you commit. Want 70B? One 24GB card cannot hold it at Q4 (~42.5GB of weights); you need a second 3090 for 48GB pooled — the server comparison covers that tier.
Performance 16GB
about $1,700 to $1,900 at MSRPstreet, Nov 2 2026: ~$2,840
CPU: Ryzen 7 9700X (8-core)
RAM: 32GB DDR5-5600
GPU: RTX 5070 Ti 16GB ($749 MSRP; $1,269.99-$1,499.99 street, Nov 2 2026)
Storage: 2TB Gen4 NVMe
Street total is the mid-range parts list in the best PC for local LLMs guide, $2,836.75 on Nov 2 2026.
📐 Bandwidth ceiling (Q4_K_M):
- • Llama 3.1 8B (~4.9GB): ~183 tok/s ceiling
- • Qwen 3 14B (~9.3GB): ~97 tok/s ceiling
- • gpt-oss:20b: MoE. OpenAI's model card lists ~3.6B active params per token, so it decodes far faster than its 20B size suggests
- RTX 5070 Ti: 896GB/s, 300W (NVIDIA spec). Twice the Sweet Spot's bandwidth on the same 16GB.
Verdict: Same models as the Sweet Spot, twice the speed ceiling. Run a dev environment and an AI coding assistant side by side; the 16GB tier is where MoE models shine. The RTX 5080 16GB ($999 MSRP, 960GB/s) adds little for LLMs over this card.
Ultimate Workstation
about $3,500 at MSRPstreet, Nov 2 2026: ~$5,800
CPU: Ryzen 9 9950X (16-core)
RAM: 64GB DDR5-5600
GPU: RTX 5090 32GB ($1,999 MSRP; Founders Edition out of stock and partner cards from $3,340, Nov 2 2026)
Storage: 4TB Gen5 NVMe
Street total is the high-end parts list in the best PC for local LLMs guide, $5,797.58 on Nov 2 2026 with a 2TB Gen4 drive (a 4TB Gen5 drive was not priced; check current price). At the $1,999 Founders Edition price the same list is ~$4,460.
📐 Bandwidth ceiling (Q4_K_M):
- • Llama 3.1 8B (~4.9GB): ~366 tok/s ceiling
- • Qwen3.6-27B (~17.4GB): ~103 tok/s ceiling
- • Qwen 2.5 Coder 32B (~19.9GB): ~90 tok/s ceiling
- • Llama 3.3 70B (Q4): loads via RAM offload, but slow. ~10GB of the ~42.5GB spills past the 32GB card, and the spilled layers run at DDR5 speed instead of the card's 1792GB/s
- RTX 5090: 1792GB/s, 575W (NVIDIA spec). Budget a 1000W or larger ATX 3.1 power supply.
Verdict: Runs Qwen3.6-27B, the current best single-GPU model, at full Q4 with long context, and the whole 32B class alongside it. Roughly double the used 3090's ceiling; only worth it if you are VRAM-bound or speed-bound today.
💡 Where These Numbers Come From
We do not own these machines. Every tokens-per-second figure on this page is an arithmetic upper bound derived from published memory bandwidth, not a benchmark run. Token generation on a local LLM is memory-bandwidth bound: every weight has to be read once per token. At Q4_K_M that is roughly 0.6GB per billion parameters, so the ceiling is simply memory bandwidth ÷ weight bytes, and the weight bytes used here are the file sizes Ollama lists for each tag (read November 2, 2026): an RTX 3090 at 936GB/s against qwen2.5-coder:32b's ~19.9GB file puts the bound near 47 tok/s. NVIDIA card VRAM, bandwidth, power and launch prices were re-read from NVIDIA's specification comparison and launch announcement on November 2, 2026; AMD's RX 9070 XT figures and Mac memory figures from Apple's tech specs were last checked in September 2026. Street prices are the retail listings read on November 2, 2026 (NVIDIA's marketplace, Walmart and Corsair) for the best PC for local LLMs guide, which keeps every source URL; used RTX 3090 prices come from resaleprices.com (1 Nov 2026) and an eBay listing seen on videocardprices.com (2 Nov 2026). VRAM fit is cross-checked against our VRAM guides. Four things pull real output below the bound:
- • Prompt processing is compute-bound, not bandwidth-bound — long prompts hit a different limit
- • The KV cache grows with context, taking VRAM and slowing each step
- • Sampling, framework overhead and quantization layout all cost throughput
- • CPU-only builds fall furthest below the bound, since DDR5 never reaches its theoretical rate
New to local AI? Start with the Windows installation guide or check which models work on your current hardware in our 8GB RAM guide. Before you buy a card, confirm it is supported in Ollama's official GPU compatibility list.
Best Models by GPU
Already own a card, or down to two? Each page lists what that exact GPU runs: LLMs, coding models and image models, with the memory each one needs.
Do You Need a Dedicated AI Server Instead of a Desktop?
Only if the machine will sit headless on your network and serve models to other devices around the clock. A server trades the desktop's flexibility for 24/7 uptime, low idle draw and more VRAM per dollar (usually via used cards), and it is a different shopping list from every build above — different case, different CPU priorities, different noise and power budget.
We keep that comparison on its own page, the best local AI server guide, rather than duplicating it here: every server tier from a ~$1,000 homelab box to dual-3090 and Mac Studio options, with the VRAM math that decides which models each one can actually hold.
Explore Hardware for 81 AI Models
Your Recommended Build
The budget build without its graphics card, priced Nov 2, 2026. Runs 7-8B models on the CPU, and the 850W supply takes a GPU later.
Specifications:
- • CPU: AMD Ryzen 5 9600X (6-core), $179.99
- • RAM: 32GB (2x16GB) DDR5-6000, $479.99
- • GPU: None (CPU only); add a 16GB card later
- • Storage: Crucial P310 1TB PCIe 4.0 NVMe, $174.99
Model Shortlist
GPU Recommendation:
What Runs on 8-12GB
Measure Performance on Your Machine
This shortlist uses budget, model size and use case. Confirm memory fit and runtime support separately. Speed depends on the exact model, quantisation, context and hardware.
- Load your chosen model and run
ollama psto check GPU or CPU placement. - Use the same prompt and context length for each comparison. Exclude the first warm-up run.
- With Ollama’s generation API, divide
eval_countbyeval_duration / 1e9for generated tokens per second. Record the model, quantisation and hardware alongside the result.
Model Memory Checker for 81 AI Models
Compare the catalogue’s memory estimates with your hardware. Leave room for the operating system, runtime and context cache; a memory match does not guarantee that a model will run. Cloud services require their provider’s API. Check the exact model file and runtime hardware support before buying.
Select Your Hardware
NVIDIA GPUs
Apple Silicon
Cloud GPUs (Monthly)
Type
gpu
Memory
12GB
Reference price
$799
Within listed memory
39/81
Airoboros 70B
Exceeds listed memoryBaichuan2 13B
Exceeds listed memoryChronos 70B
Exceeds listed memoryClaude 3 Haiku
Cloud serviceClaude 3 Opus
Cloud serviceCodeGemma 7B
Within listed memoryCodeLlama 7B
Within listed memoryCodeLlama Instruct 7B
Within listed memoryCodeLlama Python 7B
Within listed memoryCodeLlama Python 13B
Exceeds listed memoryCodestral 22B
Exceeds listed memoryCoqui TTS
Within listed memoryWhisper Large v3
Within listed memoryBark
Check memory headroomDeepSeek Coder V2 16B
Exceeds listed memoryDeepSeek Coder V2 236B
Exceeds listed memoryDolphin 2.6 Mistral 7B
Within listed memoryDolphin 2.6 Mixtral 8x7B
Exceeds listed memoryDolphin Mixtral 8x7B
Exceeds listed memoryDragon 7B
Within listed memoryFalcon 40B
Exceeds listed memoryGemma 2B
Within listed memoryGemma 7B
Within listed memoryGemma 2 2B
Within listed memoryGemma 2 9B
Within listed memoryGemma 2 27B
Exceeds listed memoryInternLM 20B
Exceeds listed memoryKoala 13B
Exceeds listed memoryLlama 2 13B
Exceeds listed memoryLlama 2 70B
Exceeds listed memoryShowing 30 of 81 models
View All Models →Need More Memory for a Downloadable Model?
Compare renting a GPU with buying hardware for your chosen model. Check the model’s licence, runtime support and memory needs first. Hosted API models stay with their provider.
Hardware Requirements by Model Category
Check Memory Before Comparing Speed
Budget for the model weights, context cache and runtime overhead. On a Mac, the operating system shares the same memory pool. For a GPU, verify how much VRAM is actually free. After loading a model, ollama ps shows whether it uses the GPU, CPU or both. Check Ollama’s memory-placement guide.
Tiny & Small (1-7B)
- RAM: 8GB minimum, 16GB recommended
- CPU: 4+ cores, modern architecture
- Storage: 50GB+ SSD space
- Check: Quantisation and space for the context cache
Medium (8-34B)
- RAM: 32GB minimum, 64GB recommended
- CPU: 8+ cores, high performance
- Storage: 100GB+ NVMe SSD
- Check: Exact model file size and available VRAM
Large & Massive (70B+)
- RAM: 64GB minimum, 128GB+ ideal
- CPU: 16+ cores, server-grade
- Storage: 200GB+ enterprise SSD
- Check: GPU residency versus system-RAM offload
Coding Models
Vision Models
Chat Models
Math Models
Best GPUs for Local AI Acceleration
NVIDIA RTX 4060 Ti 16GB
Best budget GPU for local AI with ample VRAM
- •16GB VRAM for large models
- •CUDA cores for AI acceleration
- •Runs 13B models smoothly
- •Low power consumption
NVIDIA RTX 4070 Ti
Excellent price/performance for serious AI work
- •16GB VRAM
- •Superior CUDA performance
- •Handles 30B models
- •DLSS 3 support
NVIDIA RTX 4090 24GB
Professional-grade AI workstation GPU
- •24GB VRAM for 70B models
- •Fastest inference speeds
- •Professional AI training
- •Future-proof investment
Recommended RAM Upgrades for Local AI
Corsair Vengeance 32GB Kit
Sweet spot for most local AI workloads
- •2x16GB DDR4-3600
- •Optimized for AMD & Intel
- •Run 13B models comfortably
- •Excellent heat spreaders
G.Skill Ripjaws DDR5 32GB
Latest DDR5 for newest systems
- •2x16GB DDR5-5600
- •Intel XMP 3.0
- •On-die ECC
- •Future-ready performance
Crucial 64GB DDR5 Kit
Maximum capacity for large models
- •2x32GB DDR5-6000
- •Run 70B models
- •Premium Samsung B-die
- •RGB lighting
Corsair Vengeance LPX 16GB DDR4
Affordable RAM upgrade for basic AI models
- •2x8GB DDR4-3200
- •Low profile design
- •XMP 2.0 support
- •Lifetime warranty
Pre-Built Systems for Local AI
HP Victus Gaming Desktop
Ready-to-run AI desktop under $1000
- •AMD Ryzen 7 5700G
- •16GB DDR4 RAM
- •RTX 3060 12GB
- •1TB NVMe SSD
Dell Precision 3680 Tower
Professional AI development machine
- •Intel Xeon W-2400
- •64GB ECC RAM
- •RTX 4000 Ada
- •ISV certified
Mac Mini M2 Pro
Compact powerhouse for local AI
- •M2 Pro chip
- •32GB unified memory
- •Run 30B models
- •Silent operation
Mac Studio M2 Max
Ultimate Mac for AI workloads
- •M2 Max chip
- •64GB unified memory
- •Run 70B models
- •32-core GPU
Can\'t Afford $1,000+ for Hardware? Try Cloud GPUs
Access the same powerful GPUs without the upfront cost. Perfect for testing models, occasional use, or when you need more power than your hardware provides.
Quick Cost Comparison Calculator
Cloud GPU Cost
Hardware Cost
RunPod
Affordable cloud GPUs starting at $0.2/hour
- ✓RTX 4090 at $0.74/hour
- ✓RTX 3090 at $0.44/hour
- ✓No setup required
- ✓Pay per second billing
Vast.ai
Decentralized GPU marketplace with best prices
- ✓RTX 4090 from $0.40/hour
- ✓50% cheaper than AWS
- ✓Global availability
- ✓Instant deployment
Lambda Labs
Professional GPU cloud for AI/ML teams
- ✓A100 80GB available
- ✓Persistent storage
- ✓Jupyter notebooks
- ✓Team collaboration
Paperspace
User-friendly GPU cloud with free tier
- ✓Free GPU tier available
- ✓One-click templates
- ✓AutoML tools
- ✓Gradient notebooks
Cloud vs Local: Quick Comparison
| Aspect | Cloud GPU | Local Hardware |
|---|---|---|
| Initial Cost | ✓ $0 upfront | ✗ $800-15,000 |
| Scalability | ✓ Instant scaling | ✗ Fixed capacity |
| Maintenance | ✓ Zero maintenance | ✗ Your responsibility |
| Privacy | ⚠ Data leaves premises | ✓ 100% local |
| Latency | ⚠ Network dependent | ✓ No network latency |
| 24/7 Usage | ✗ Expensive | ✓ Fixed cost |
Start with Cloud, Upgrade to Local Later
The smart approach: Test models and learn on cloud GPUs for $20-50/month. Once you know exactly what you need, invest in the right hardware.
🎓 Learn How to Use Cloud GPUs
Step-by-step tutorials showing exactly how to run AI models on cloud GPUs. Start in 5 minutes for just $10.
Reference Build Guides
Detailed component lists optimized for different model sizes and use cases. These totals are the page's earlier costing, last re-checked in September 2026, and they have not been re-priced since. GPU and DDR5 prices kept climbing: on November 2, 2026 a 32GB DDR5-6000 kit was $479.99 and a 64GB kit $833.63 at Walmart, and the cheapest in-stock RTX 5090 was $3,340. Treat every total below as a floor and check current prices on the day you buy; the five builds above carry November 2026 street totals.
Student Build
$799- • AMD Ryzen 5 5600 (6-core)
- • 16GB DDR4-3200 RAM
- • 500GB NVMe SSD
- • Used RTX 3060 12GB
- • 550W PSU, mATX case
Developer Build
$1,899- • AMD Ryzen 7 7700X (8-core)
- • 32GB DDR5-5600 RAM
- • 1TB Samsung 980 PRO
- • RTX 4070 12GB
- • 750W Gold PSU
AI Researcher
$3,499- • Intel i9-13900K (24-core)
- • 64GB DDR5-6000 RAM
- • 2TB Samsung 990 PRO
- • RTX 4080 16GB
- • 1000W Platinum PSU
Mac mini M5 Pro
From $1,699Apple Store, Nov 2 2026
- • M5 Pro chip (15-core CPU)
- • 24GB unified memory (307GB/s, Apple spec)
- • 512GB SSD
- • 16-core GPU
- • Silent, tiny, ~no maintenance
Pro Workstation
$5,999- • AMD Threadripper PRO
- • 128GB ECC RAM
- • 4TB NVMe RAID
- • RTX 4090 24GB
- • 1600W Redundant PSU
Enterprise Server
$10K+- • Dual EPYC or Xeon
- • 256GB+ ECC RAM
- • 8TB Enterprise SSD
- • Dual RTX 4090/5090 or A6000
- • 4U Rackmount
RTX 5090 Flagship
$5,500+- • AMD Ryzen 9 9950X (16-core)
- • 64GB DDR5-6000
- • 2TB Gen5 NVMe
- • RTX 5090 32GB
- • 1200W ATX 3.1 PSU
Throughput Ceilings by Build Tier
These are not benchmark results. Each figure is the arithmetic upper bound for Q4_K_M decoding on that machine — published memory bandwidth divided by the weight bytes moved per token. Treat them as the speed a tier cannot exceed, and expect real output somewhere below.
Arithmetic Ceiling — Llama 3.1 8B at Q4_K_M (~4.9GB of weights)
| Hardware Configuration | Model (Q4_K_M) | Memory Bandwidth | Weights Read Per Token | Arithmetic Ceiling |
|---|---|---|---|---|
| Budget Build (Ryzen 5 9600X, 32GB DDR5, CPU only) | Llama 3.1 8B | ~90GB/s (dual-channel DDR5-5600) | ~4.9GB | ~18 tok/s |
| Sweet Spot Build (Ryzen 5 9600X, 32GB, RTX 5060 Ti 16GB) | Llama 3.1 8B | 448GB/s | ~4.9GB | ~91 tok/s |
| Sweet Spot Build (Ryzen 5 9600X, 32GB, RTX 5060 Ti 16GB) | Qwen 2.5 Coder 14B | 448GB/s | ~9.0GB | ~50 tok/s |
| Performance 16GB (Ryzen 7 9700X, 32GB, RTX 5070 Ti 16GB) | Qwen 3 14B | 896GB/s | ~9.3GB | ~97 tok/s |
| Ultimate Workstation (Ryzen 9 9950X, 64GB, RTX 5090 32GB) | Qwen3.6-27B | 1792GB/s | ~17.4GB | ~103 tok/s |
| AI Server (Ryzen 7 7700, 64GB, dual used RTX 3090 — 48GB pooled) | Llama 3.3 70B | 936GB/s per card (layers run serially, so one card's bandwidth sets the pace) | ~42.5GB | ~22 tok/s |
* Ceilings, not benchmarks: bandwidth ÷ weight bytes, using NVIDIA's and AMD's published bandwidth figures and the Q4_K_M file sizes Ollama lists for each model. Real throughput is lower once prefill, KV cache growth and sampling overhead are included. The 70B row is the dual-3090 tier, not a single card: a 70B at Q4_K_M is ~42.5GB of weights, so 48GB of pooled VRAM is the floor — see the local AI server comparison for the full breakdown.
GPU Memory Bandwidth Comparison
Bandwidth is the number that decides decoding speed, so it is the honest way to rank cards. The ceiling column is bandwidth ÷ ~9.3GB, the size Ollama lists for qwen3:14b at Q4_K_M (the weights a 14B model moves per token).
| GPU | VRAM | Memory Bandwidth | 14B Ceiling | Largest Dense Model at Q4 |
|---|---|---|---|---|
| RTX 3060 | 12GB | 360GB/s | ~39 tok/s | 14B class |
| RTX 4070 | 12GB | 504GB/s | ~54 tok/s | 14B class |
| RTX 5060 Ti (16GB) | 16GB | 448GB/s | ~48 tok/s | 14B class, 20B MoE |
| RX 9070 XT (AMD) | 16GB | 640GB/s | ~69 tok/s | 14B class, 20B MoE |
| RTX 4080 Super | 16GB | 736GB/s | ~79 tok/s | 20-24B |
| RTX 5070 Ti | 16GB | 896GB/s | ~97 tok/s | 20-24B |
| RTX 5080 | 16GB | 960GB/s | ~103 tok/s | 20-24B |
| RTX 3090 (used) | 24GB | 936GB/s | ~101 tok/s | 27-32B |
| RTX 4090 | 24GB | 1008GB/s | ~109 tok/s | 27-32B |
| RTX 5090 | 32GB | 1792GB/s | ~193 tok/s | 32B with long context; ~49B is tight |
Bandwidth and VRAM are NVIDIA's and AMD's published specs. Ceilings are arithmetic upper bounds, not measured throughput. Street prices are running far above MSRP right now — see why GPU prices are so high.
Hardware FAQ
Do I need a GPU for local AI?
Not necessarily. Modern CPUs can run smaller models (3B-8B) effectively. However, a GPU provides 2-5x speed improvements and enables running larger models more efficiently. If you plan to use AI regularly or work with larger models, a GPU is highly recommended.
How much RAM do I really need?
RAM is crucial for local AI. As a rule of thumb: model size + 4-8GB for the operating system. For an 8B model (~5GB), you need at least 12GB RAM, but 16GB+ is recommended for smooth operation. For 70B models, you need 64GB+ RAM.
Should I build a desktop or a dedicated AI server?
Build a desktop if you are the only user and you want the machine to do other things too — every build on this page fits that case. Go headless-server only when models need to be available on your network 24/7, which changes the CPU, case, noise and power calculus entirely. Tier-by-tier server options (homelab, dual-3090, Mac Studio) are compared on the best local ai server page, with the single-box parts list in the $1,500 AI server build guide.
Is Apple Silicon (M-series) good for AI?
Yes! Apple Silicon offers excellent AI performance with unified memory architecture. A base Mac Mini M4 (16GB) handles 7-8B models, M4 Pro/Max configurations run the 14B-70B range depending on memory, and unified memory means the model shares one big pool instead of fighting a VRAM ceiling. See the Apple M4 local AI guide for chip-by-chip picks.
Can I upgrade my existing computer?
Often yes! The most impactful upgrades are usually RAM (if your motherboard supports more) and adding a GPU. However, very old CPUs (pre-2018) may become bottlenecks. Check your motherboard specifications for RAM and GPU compatibility.
Which models can I run with my hardware?
Start by checking the Local AI Models directory to filter by parameters, modality, and context window that match your build. If you're on a lean system, jump into the 8GB optimization guide for hand-picked quantized models before upgrading to larger tiers.
Was this helpful?
Get Hardware Updates & Deals
Get the latest hardware recommendations, price and availability updates, and deals delivered weekly, written for people running models at home.
Build Real AI on Your Machine
RAG, agents, NLP, vision, and MLOps - chapters across 25 courses that take you from reading about AI to building AI.
Written by
Barsha Mohapatra
Founder and Editor, Local AI Master · Founder, Little AI Master
Founder of Little AI Master and the editor behind every course and guide on Local AI Master. Years of tutoring students taught her where AI education actually fails.
Related Guides
Continue your local AI journey with these comprehensive guides