# localaiprices

> Live prices (USD and EUR), stock, and measured tokens/sec per $1,000 and per €1,000 for local-AI hardware: AMD Strix Halo mini PCs, NVIDIA DGX Spark/GB10 boxes, Apple Mac Studio, and RTX GPUs. Updated on a schedule; all data is free static JSON.

## API
- [All machines with prices and value scores](https://localaiprices.com/api/v1/skus.json): JSON, sorted by tok/s per $1k
- [Latest price observations](https://localaiprices.com/api/v1/prices/latest.json)
- [Benchmark sources](https://localaiprices.com/api/v1/benchmarks.json): llama.cpp community threads plus llm-speed.com (CC BY 4.0)
- [Cloud GPU hourly prices](https://localaiprices.com/api/v1/cloud.json)

## Pages
- [Comparisons](https://localaiprices.com/compare/): 18 head-to-head pages ([JSON](https://localaiprices.com/api/v1/compare.json))
- [Models](https://localaiprices.com/models/): memory, fit and measured tok/s for 16 popular open models ([JSON](https://localaiprices.com/api/v1/models.json), [benchmarks JSON](https://localaiprices.com/api/v1/model_benchmarks.json))
- [Ranking](https://localaiprices.com/)
- [Rent instead](https://localaiprices.com/rent/)
- [Benchmarks](https://localaiprices.com/benchmarks/)

## Notes
- Small-model metric: llama.cpp llama-bench, Llama-2 7B Q4_0, tg128 (text generation tok/s); best of FA on/off as posted
- Large-model metric: gpt-oss-120b MXFP4 (59 GiB weights), text generation tok/s at empty context, single request (Needs about 59 GiB for weights plus context and OS, so devices with 64GB or less are marked 'doesn't fit'.)
- kind=gpu-card items are graphics cards only and need a host PC
- Fields ending in _converted=true were converted at the ECB EUR/USD reference rate (1.1253, 2026-10-16).

## Current ranking
- Mac Studio M5 Max 40-core GPU, 128GB / 512GB: $5099 / €5909, 119.08 tok/s, 23.35 tok/s per $1k small; large model: 92.8, 18.2 per $1k [machine] (https://localaiprices.com/p/mac-studio-m5-max-128/)
- Mac Studio M5 Ultra 64-core GPU, 96GB / 1TB: $5499 / €6699, no bench tok/s, - tok/s per $1k small; large model: 97.2, 17.68 per $1k [machine] (https://localaiprices.com/p/mac-studio-m5-ultra-96/)
- Framework Desktop 128GB (DIY): $3449 / €3889, 50.95 tok/s, 14.77 tok/s per $1k small; large model: 55.57, 16.11 per $1k [machine] (https://localaiprices.com/p/framework-desktop-128/)
- GMKtec EVO-X2 128GB / 1TB: $3499.99 / €3299.99, 50.95 tok/s, 14.56 tok/s per $1k small; large model: 55.57, 15.88 per $1k [machine] (https://localaiprices.com/p/gmktec-evo-x2-128-1tb/)
- GMKtec EVO-X2 128GB / 2TB: $3649.99 / €3449.99, 50.95 tok/s, 13.96 tok/s per $1k small; large model: 55.57, 15.22 per $1k [machine] (https://localaiprices.com/p/gmktec-evo-x2-128-2tb/)
- Geekom A9 Mega 128GB / 2TB: $3799.0 / €3999.0, 50.95 tok/s, 13.41 tok/s per $1k small; large model: 55.57, 14.63 per $1k [machine] (https://localaiprices.com/p/geekom-a9-mega-128-2tb/)
- GMKtec EVO-X3 128GB / 2TB: $3799.99 / €3549.99, 50.95 tok/s, 13.41 tok/s per $1k small; large model: 55.57, 14.62 per $1k [machine] (https://localaiprices.com/p/gmktec-evo-x3-128-2tb/)
- Minisforum MS-S1 MAX 128GB / 2TB: $3879.0 / €3447.08, 50.95 tok/s, 13.13 tok/s per $1k small; large model: 55.57, 14.33 per $1k [machine] (https://localaiprices.com/p/minisforum-ms-s1-max-128/)
- MacBook Pro 16" M5 Max 40-core GPU, 128GB / 2TB: $6999 / €7999, 119.08 tok/s, 17.01 tok/s per $1k small; large model: 92.8, 13.26 per $1k [machine] (https://localaiprices.com/p/macbook-pro-16-m5-max-128/)
- Beelink GTR9 Pro 128GB / 2TB: $4349.0 / €3864.75, 50.95 tok/s, 11.72 tok/s per $1k small; large model: 55.57, 12.78 per $1k [machine] (https://localaiprices.com/p/beelink-gtr9-pro-128/)
- NVIDIA DGX Spark (Founders Edition): $4699 / €4175.78, 57.21 tok/s, 12.17 tok/s per $1k small; large model: 58.72, 12.5 per $1k [machine] (https://localaiprices.com/p/nvidia-dgx-spark/)
- Dell Pro Max / Pro Precision with GB10, 128GB / 2TB: $5903 / €5245.71, 57.21 tok/s, 9.69 tok/s per $1k small; large model: 58.72, 9.95 per $1k [machine] (https://localaiprices.com/p/dell-pro-max-gb10/)
- ASUS Ascent GX10 (GB10): $5999 / €5331.02, 57.21 tok/s, 9.54 tok/s per $1k small; large model: 58.72, 9.79 per $1k [machine] (https://localaiprices.com/p/asus-ascent-gx10/)
- GMKtec EVO-X2 64GB / 1TB: $2199.99 / €1999.99, 50.95 tok/s, 23.16 tok/s per $1k small; large model: doesn't fit in memory, - per $1k [machine] (https://localaiprices.com/p/gmktec-evo-x2-64-1tb/)
- Minisforum MS-S1 MAX 64GB / 2TB: $2599.0 / €2309.61, 50.95 tok/s, 19.6 tok/s per $1k small; large model: doesn't fit in memory, - per $1k [machine] (https://localaiprices.com/p/minisforum-ms-s1-max-64/)
- Geekom A9 Mega 64GB / 2TB (Ryzen AI Max+ 388): $2599.0 / €2499.0, no bench tok/s, - tok/s per $1k small; large model: doesn't fit in memory, - per $1k [machine] (https://localaiprices.com/p/geekom-a9-mega-64-2tb/)
- Framework Desktop 64GB (DIY): $2485.79 / €2209, 50.95 tok/s, 20.5 tok/s per $1k small; large model: doesn't fit in memory, - per $1k [machine] (https://localaiprices.com/p/framework-desktop-64/)
- HP Z2 Mini G1a (Ryzen AI Max+ PRO 395): $n/a / €n/a, 50.95 tok/s, - tok/s per $1k small; large model: 55.57, - per $1k [machine] (https://localaiprices.com/p/hp-z2-mini-g1a/)
- MSI EdgeXpert MS-C931 (GB10): $n/a / €n/a, 57.21 tok/s, - tok/s per $1k small; large model: 58.72, - per $1k [machine] (https://localaiprices.com/p/msi-edgexpert/)
- Lenovo ThinkStation PGX (GB10): $n/a / €n/a, 57.21 tok/s, - tok/s per $1k small; large model: 58.72, - per $1k [machine] (https://localaiprices.com/p/lenovo-thinkstation-pgx/)
- Mac Studio M5 Ultra 80-core GPU, 256GB / 1TB: $10799 / €12529, 179.1 tok/s, 16.58 tok/s per $1k small; large model: no public result yet, - per $1k [machine] (https://localaiprices.com/p/mac-studio-m5-ultra-80-256/)
- Mac mini M5 Pro 16-core GPU, 64GB / 512GB: $2699 / €3129, 67.59 tok/s, 25.04 tok/s per $1k small; large model: doesn't fit in memory, - per $1k [machine] (https://localaiprices.com/p/mac-mini-m5-pro-16/)
- Mac mini M5 Pro 20-core GPU, 64GB / 512GB: $2899 / €3349, no bench tok/s, - tok/s per $1k small; large model: doesn't fit in memory, - per $1k [machine] (https://localaiprices.com/p/mac-mini-m5-pro-20/)
- NVIDIA RTX 5090 32GB (new): $8213.56 / €7299.0, 314.0 tok/s, 38.23 tok/s per $1k small; large model: doesn't fit in memory, - per $1k [gpu-card] (https://localaiprices.com/p/rtx-5090/)
- NVIDIA RTX 5080 16GB (new): $1895.01 / €1684.0, 184.68 tok/s, 97.46 tok/s per $1k small; large model: doesn't fit in memory, - per $1k [gpu-card] (https://localaiprices.com/p/rtx-5080/)
- NVIDIA RTX 5070 Ti 16GB (new): $1400.04 / €1244.15, 182.43 tok/s, 130.3 tok/s per $1k small; large model: doesn't fit in memory, - per $1k [gpu-card] (https://localaiprices.com/p/rtx-5070-ti/)
- NVIDIA RTX PRO 6000 Blackwell Max-Q 96GB (new): $20086.6 / €17850.0, 287.29 tok/s, 14.3 tok/s per $1k small; large model: no public result yet, - per $1k [gpu-card] (https://localaiprices.com/p/rtx-pro-6000-maxq/)
- NVIDIA RTX 4090 24GB (used): $n/a / €n/a, 188.96 tok/s, - tok/s per $1k small; large model: doesn't fit in memory, - per $1k [gpu-card] (https://localaiprices.com/p/rtx-4090-used/)
- NVIDIA RTX 3090 24GB (used): $700 / €622.06, 177.79 tok/s, 253.99 tok/s per $1k small; large model: doesn't fit in memory, - per $1k [gpu-card] (https://localaiprices.com/p/rtx-3090-used/)