Best home AI rig under €5,000
Only builds whose total cost is stated in the source (converted to € at the ECB rate) are listed. Reported costs are what the builder said they paid.
Builds with a source-reported total cost
| Build | GPU(s) | VRAM GB | Reported cost | Measured tok/s |
|---|---|---|---|---|
| DGX Spark (GB10): Qwen3.6 35B-A3B NVFP4 at 113 tok/s on vLLM with MTP | Nvidia GB10 (unified) | 128 | $3,999 (≈€3,568) | Qwen3.6 35B-A3B: 113 Qwen3.6 35B-A3B: 86 |
| Minisforum MS-S1 MAX (Strix Halo, 128 GB): Qwen3.6 35B-A3B at 62 tok/s on llama.cpp Vulkan | AMD Ryzen AI Max+ 395 (Radeon 8060S iGPU) | 128 | $3,799 (≈€3,390) | Qwen3.6 35B-A3B: 62 Qwen3.6 35B-A3B: 54 Qwen3.6 35B-A3B: 51 Qwen3.6 35B-A3B: 93 |
| Dual used RTX 3090, ~2000 EUR total ⚠ unchecked | 2x RTX 3090 (used) | 48 | €2,000 (≈$2,242) | Llama 3.3 70B: 20 |
| Mac Studio 64 GB ⚠ unchecked | Apple unified memory | 64 | $4,000 (≈€3,569) | Llama 3.3 70B: 11 |
| $2K EPYC 64-core, 512 GB DDR4, CPU-only | None (CPU inference) | $2,000 (≈€1,784) | DeepSeek R1 671B: 4.31 | |
| 4x RTX 3090 homelab on Threadripper 2920X ⚠ unchecked | 4x RTX 3090 (~$625 each, used) | 96 | $3,300 (≈€2,944) | no speed data |
Builds whose GPUs cost under €5,000 today (estimated)
These sources don't state a total cost. The figure is our estimated GPU cost only, from the lowest new retail price on billiger.de, approx. price as of 2026-10-16. CPU, RAM, PSU, case and storage are not included, and used GPUs often cost less. Each build page links the exact price source.
| Build | GPU(s) | VRAM GB | Est. GPU cost | Measured tok/s |
|---|---|---|---|---|
| Intel Core Ultra 265K + RTX 4090 | RTX 4090 | 24 | ≈€3,955 GPU only, est. | gpt-oss 20B: 279.65 DeepSeek R1 Distill 14B: 92.49 Llama 3.2 3B: 334.67 |
| RTX 4090: 2026 models (llamabench community) | RTX 4090 | 24 | ≈€3,955 GPU only, est. | Qwen3.6 27B: 47.0 Gemma 4 31B: 43.8 Gemma 4 26B-A4B: 164 Gemma 4 12B: 98.1 gpt-oss 20B: 203 Qwen3.5 9B: 137 LFM2.5 24B-A2B: 261-339 Nemotron 3.5 Bolt 74B-A9B: 108 |
| Ryzen 9 7900X + RTX 4090 | RTX 4090 | 24 | ≈€3,955 GPU only, est. | Llama 3.2 3B: 237.05 Llama 3.1 70B: 3.1 |
| Dual RTX 5070 Ti on Ryzen 9 9900X (capy-flash maintainer rig) | 2x RTX 5070 Ti 16 GB | 32 | ≈€2,375 GPU only, est. | Qwen3.8 Flash-Next: 182 Qwen3.8 Flash-Next: 155 Qwen3.8 27B: 139 Qwen3.8 27B: 230 |
| Single RTX 5080 16GB | RTX 5080 | 16 | ≈€1,599 GPU only, est. | gpt-oss 20B: 204.85 |
| Ryzen 9 7950X + RTX 4080 | RTX 4080 | 16 | ≈€1,490 GPU only, est. | Llama 3.2 3B: 204.45 |
| Ryzen 9 7950X + RTX 4070 Ti Super | RTX 4070 Ti Super | 16 | ≈€1,490 GPU only, est. | Llama 3.2 3B: 198.95 |
| Intel Core Ultra 265K + Radeon AI Pro R9700 | AMD Radeon AI Pro R9700 | 32 | ≈€1,718 GPU only, est. | gpt-oss 20B: 163.26 DeepSeek R1 Distill 14B: 53.72 Llama 3.2 3B: 195.91 |
| Single RTX 5070 Ti 16GB | RTX 5070 Ti | 16 | ≈€1,188 GPU only, est. | gpt-oss 20B: 189.45 |
| Single RTX 4090 (llama.cpp CUDA scoreboard) | RTX 4090 24GB | 24 | ≈€3,955 GPU only, est. | Llama 2 7B: 188.96 |
| RTX 4090 (llm-speed signed run, Oct 2026): Flint 40B-A2.5B at 169-179 tok/s on llama.cpp | RTX 4090 | 24 | ≈€3,955 GPU only, est. | Flint 40B-A2.5B: 179.4 Flint 40B-A2.5B: 169.2 |
| RTX 4090 (llm-speed signed runs, Oct 2026): Qwen3.6 35B-A3B 161 tok/s, Gemma 4 12B QAT 98 tok/s | RTX 4090 | 24 | ≈€3,955 GPU only, est. | Qwen3.6 35B-A3B: 160.6 Gemma 4 12B: 97.7 GLM 4.7 Flash: 164.9 |
| Intel Core Ultra 265K + RTX A4000 | RTX A4000 | 16 | ≈€1,050 GPU only, est. | gpt-oss 20B: 93.55 Llama 3.2 3B: 149.44 |
| Dual RTX 5080 on EPYC 7F52 (capy-flash user report) | 2x RTX 5080 16 GB | 32 | ≈€3,198 GPU only, est. | Qwen3.8 Flash-Next: 141.3 Qwen3.8 Flash-Next: 144.4 |
| Ryzen 9 9950X + Radeon RX 7900 XT | AMD Radeon RX 7900 XT | 20 | ≈€869 GPU only, est. | Llama 3.2 3B: 131.2 |
| RTX 3060 12GB: 2026 models (llamabench community) | RTX 3060 | 12 | ≈€429 GPU only, est. | Gemma 4 12B: 38.5 LFM2.5 24B-A2B: 125 Qwen3.8 27B: 5.2-6.7 |
| Intel Core Ultra 265K + RTX 3060 | RTX 3060 | 12 | ≈€429 GPU only, est. | gpt-oss 20B: 83.96 DeepSeek R1 Distill 14B: 29.77 Llama 3.2 3B: 122.85 |
| Radeon AI Pro R9700 (one card): Qwen3.6 35B-A3B from 4K to 256K context (AMDai lab rig) | AMD Radeon AI Pro R9700 (1 of 2 cards in the rig used) | 32 | ≈€1,718 GPU only, est. | Qwen3.6 35B-A3B: 102.39 Qwen3.6 35B-A3B: 90.52 Qwen3.6 35B-A3B: 66.43 Qwen3.6 35B-A3B: 44.4 Qwen3.6 35B-A3B: 109.73 Qwen3.6 35B-A3B: 117.69 |
| Single RTX 5060 Ti 16GB | RTX 5060 Ti 16GB | 16 | ≈€739 GPU only, est. | gpt-oss 20B: 111.51 |
| RTX 5070 Ti: 2026 models (llamabench community) | RTX 5070 Ti | 16 | ≈€1,188 GPU only, est. | Gemma 4 12B: 87.6 |
| Radeon AI Pro R9700: Qwen3.6 27B on Vulkan, HIP and Ollama, with and without MTP | AMD Radeon AI Pro R9700 (1 of 2 cards in the rig used) | 32 | ≈€1,718 GPU only, est. | Qwen3.6 27B: 66.0 Qwen3.6 27B: 31.9 Qwen3.6 27B: 27.1 Qwen3.6 27B: 26.7 |
| Radeon AI Pro R9700: Qwen3.8 27B on llama.cpp, Ollama and vLLM (Vadi Taslim) | AMD Radeon AI Pro R9700 (1 of 2 cards in the rig used) | 32 | ≈€1,718 GPU only, est. | Qwen3.8 27B: 57 Qwen3.8 27B: 30 Qwen3.8 27B: 22 |
| RTX 4090 desktop (speed comparison vs Macs) | RTX 4090 24GB | 24 | ≈€3,955 GPU only, est. | Llama 3.1 8B: 52.99 Mistral Small 22B: 26.72 |
| i7-12700K + RTX 5060 Ti 16GB + RTX 3060 12GB: Qwen3.8 27B from 21 to 51 tok/s (split, MTP, single-GPU) | RTX 5060 Ti 16GB + RTX 3060 12GB | 28 | ≈€1,168 GPU only, est. | Qwen3.8 27B: 21.2 Qwen3.8 27B: 37.6 Qwen3.8 27B: 28.0 Qwen3.8 27B: 51.4 |
| RTX A4000 16GB (llm-speed signed run, Oct 2026): Muse Glimmer 30B UD-IQ3_XXS at 21.4 tok/s | RTX A4000 | 16 | ≈€1,050 GPU only, est. | Muse Glimmer 30B: 21.42 |
96 more builds have measured speeds but no cost in their source; browse them on the main table or by model. Know what your rig cost? Submit it.
Rent before you buy: try this model on a cloud GPU for a few dollars first: Runpod (referral link; sign up with Google to get the new-user bonus)