Explore builds
STARTER BUILD / RigAtlas starter

Ryzen 7 7800X3D + RTX 4070 PC Build.

Reference imageImages are for reference only. Refer to the linked store for actual product images.
ABOUT THIS BUILD

An eight-core 3D V-Cache CPU and a 12 GB graphics card, ready to adapt to your games and workload. Check thermal performance, BIOS support and current pricing. Images are for reference only. Refer to the linked store for actual product images.

8 componentsBudget in USD
CONFIGURATION

AM5 with more headroom

8 / 8 core parts
COMPONENTYOUR SELECTIONRECORDED PRICE
Processor
AM5 · DDR5
Checking price…
Not recorded
Motherboard
AM5 · DDR5 · ATX
Checking price…
Not recorded
Graphics card
12 GB GDDR6X
Checking price…
Not recorded
Memory
DDR5 · UDIMM
Checking price…
Not recorded
Storage
M.2 2280
Checking price…
Not recorded
Power supply
ATX
Checking price…
Not recorded
CPU cooler
Dual-tower air cooler
Checking price…
Not recorded
Case
Mid-tower
Checking price…
Not recorded
Recorded prices are entered by you in USD. Retailer prices, tax and delivery may differ.
AI & YOUR BUILD

What AI work can this PC handle?

Explore cloud use, local language models and image generation with the parts you selected. Memory estimates are not speed benchmarks.

System memory32 GB
Dedicated graphics memory12 GB
Selected graphics cardGeForce RTX 4070 VENTUS 2X 12G OC

Cloud AI

Inference runs on the service’s servers.

A larger local GPU is not required just to chat with a cloud model or request cloud-generated images. Check the service account, availability in your region, internet connection and supported browser or client.

Local editing, 3D work and running several apps still use your PC’s resources. This does not guarantee a service’s response time.

How cloud inference works

Local image generation

Choose the exact model and workflow first.

Check the exact GPU, driver and CUDA support in your chosen application.

Image size, batch size, model components and extra nodes change memory use. CPU offloading can reduce VRAM demand but may be slow. Video generation, training and fine-tuning require separate assessments.

ComfyUI hardware requirementsModel memory and offloading

Local language models

Change the assumptions to explore the memory tradeoff. These are model sizes, not recommendations for a specific model.

Weights may fit; full inference still needs verification.

Theoretical weight storage only: 3.7 GiB · 8B / 4-bit

Actual use is higher: quantization metadata, unquantized tensors, KV cache, compute buffers and the runtime also need memory. Total installed memory is not the same as free memory.

Weight storage scenarios at the selected precision
ParametersWeight lower boundCompared with selected VRAM
3B1.4 GiBWeights only; overhead not included
8B3.7 GiBWeights only; overhead not included
14B6.5 GiBWeights only; overhead not included
32B14.9 GiBNot enough for weights alone
70B32.6 GiBNot enough for weights alone
  • Check the exact GPU, driver and CUDA support in your chosen application.
  • Longer context increases KV-cache demand. A weights-only fit does not establish whether your selected context or concurrent requests will fit.
  • Check CPU instruction support, the exact OS version, drivers and available SSD space for model downloads. LM Studio on Windows x64 requires AVX2; this catalog does not verify instruction sets.
Method, limits and how to verify

Weight lower bound = total parameters × bits ÷ 8. B means one billion parameters; GiB means 2³⁰ bytes. All parameters are counted, including inactive experts in a mixture-of-experts model. Nominal catalog RAM and VRAM capacities are used as GiB for this planning comparison.

Shared system memory is not added to dedicated VRAM. RAM kit capacity is counted once. Already-owned parts still count as installed hardware. No throughput or “smooth performance” claim is made.

After assembly, use the actual model file and context setting with the runtime’s resource estimator, then measure response speed and memory use in your workload.

LM Studio resource estimation

Weight estimates · Context and memory · GPU runtime support · Application requirements

Selected GPU manufacturer specifications · 2026-09-09

AI documentation checked: 2026-09-19
THE CHOICES BEHIND THIS PARTS LIST

Who this build is for

A Ryzen 7 7800X3D and RTX 4070 configuration retained for comparing an older X3D platform with newer complete baskets.

Why these components are together

The X3D processor makes gaming CPU behavior part of the choice, while the 12 GB RTX 4070 has separate rendering and software limits. Use exact workload evidence for both.

The 32 GB DDR5 kit and 1 TB SSD keep capacity moderate. More CPU cache does not replace system memory or graphics memory when a project needs those resources.

The NH-D15 and 4000D AIRFLOW require a joint fan-height and memory-clearance check. Noctua's current height information differs from some older material, so identify the actual cooler and installed fan position.

Tradeoffs and checks before ordering

  • A raised front fan can exceed a case limit even when the bare cooler specification appears close.
  • Keep the current compatibility review unresolved until the actual dimensions or a manufacturer-confirmed arrangement settles it.
  • Compare the delivered cost against newer X3D and GPU options instead of treating an older recommended list as permanently best value.

What to change first

If the cooler arrangement cannot be confirmed, choose a documented-fitting cooling alternative before ordering. For a performance upgrade, first measure the workload and compare a complete change rather than replacing the CPU by generation alone.

This is an editorial parts-list explanation. The planning target is not a current retailer total, and this configuration has not been benchmarked by RigAtlas.

Component specification sources