Local AI hardware library

MacBook Pro M5 Max

Choose a memory configuration. Model fit and price are calculated per variant.

2 configurations

36GB unified memory

Apple36GB unified$3,499ollama, llama.cpp

Entry M5 Max configuration with high bandwidth and GPU throughput, but less model capacity than higher-memory M5 Pro SKUs.

Chat 100Code 98RAG 98

128GB unified memory

Apple128GB unified$6,499ollama, llama.cpp

Maximum-memory M5 Max MacBook Pro with 614 GB/s bandwidth for large models and demanding local AI workflows.

MacBook Pro M5 Pro

Choose a memory configuration. Model fit and price are calculated per variant.

2 configurations

24GB unified memory

Apple24GB unified$2,499ollama, llama.cpp

Entry M5 Pro MacBook Pro with twice the base M5 memory bandwidth, but a 24GB capacity ceiling for model fit.

64GB unified memory

Apple64GB unified$3,899ollama, llama.cpp

Maximum-memory M5 Pro MacBook Pro, balancing 64GB capacity with 307 GB/s bandwidth for portable local AI.

MacBook Air M5

Choose a memory configuration. Model fit and price are calculated per variant.

2 configurations

16GB unified memory

Apple16GB unified$1,099ollama, llama.cpp

Fanless M5 MacBook Air for portable local AI. The 16GB SKU is best suited to compact models and conservative context sizes.

32GB unified memory

Apple32GB unified$1,499ollama, llama.cpp

Highest-memory M5 MacBook Air configuration. Its 32GB pool expands model fit, but the fanless chassis limits sustained heavy workloads.

HP EliteBook X G2a Ryzen AI 400 64GB

HP / AMD64GB unified$1,899Runtime varies

Ryzen AI 400 business laptop with up to 64GB soldered memory. Local model acceleration depends on ROCm, Vulkan, DirectML, and application support.

Chat 84Code 82RAG 82

ASUS Zenbook A16 Snapdragon X2 Elite Extreme 48GB

ASUS / Qualcomm48GB unified$1,999Runtime varies

A real Snapdragon X2 Elite Extreme retail system. Runtime compatibility on Windows Arm should be checked before choosing a specific local AI tool.

Qualcomm / Windows OEMs128GB unified$2,299Runtime varies

Chat 78Code 76RAG 76

Snapdragon X2 Elite Extreme 128GB AI PC

AI PCannounced

Platform-level profile for high-memory Snapdragon X2 Elite Extreme Windows AI PCs. Real performance depends on the OEM device, memory configuration, thermals, and runtime support.

NVIDIA128GB unified$3,999ollama, llama.cpp

NVIDIA DGX Spark 128GB

desktopCurrent

Compact Grace Blackwell desktop system for local agents, large-model inference, and AI development.

MacBook Pro M5

Choose a memory configuration. Model fit and price are calculated per variant.

2 configurations

16GB unified memory

Apple16GB unified$1,999ollama, llama.cpp

Actively cooled base M5 MacBook Pro. Memory, rather than compute, is the main limitation of this 16GB configuration.

32GB unified memory

Apple32GB unified$2,599ollama, llama.cpp

Base M5 MacBook Pro with its maximum 32GB unified-memory configuration.

Mac Studio M3 Ultra 192GB

Apple192GB unified$10,799ollama, llama.cpp

Mac Studio M3 Ultra 192GB profile for local AI model planning.

NVIDIA RTX 6000 Blackwell 96GB

NVIDIA96GB VRAM$8,999ollama, llama.cpp

NVIDIA RTX 6000 Blackwell 96GB profile for local AI model planning.

AMD Radeon RX 9070 XT 16GB

AMD16GB VRAM$599Runtime varies

AMD Radeon RX 9070 XT 16GB profile for local AI model planning.

Chat 80Code 78RAG 78

Mac Studio M4 Max 64GB

Apple64GB unified$3,799ollama, llama.cpp

Mac Studio M4 Max 64GB profile for local AI model planning.

Ryzen AI Max+ 395 128GB Mini PC

AMD / Mini PC OEMs128GB unified$1,999Runtime varies

Ryzen AI Max+ 395 128GB Mini PC profile for local AI model planning.

MacBook Air M4 16GB

Apple16GB unified$1,299ollama, llama.cpp

MacBook Air M4 16GB profile for local AI model planning.

NVIDIA RTX 5070 Ti 16GB

NVIDIA16GB VRAM$749ollama, llama.cpp

NVIDIA RTX 5070 Ti 16GB profile for local AI model planning.

Vision 89Image 89Chat 80

Framework Desktop AI Max 128GB

Framework / AMD128GB unified$1,999Runtime varies

Framework Desktop AI Max 128GB profile for local AI model planning.

NVIDIA RTX 5090 32GB

NVIDIA32GB VRAM$1,999ollama, llama.cpp

NVIDIA RTX 5090 32GB profile for local AI model planning.

Vision 93Image 93Chat 88

NVIDIA RTX 5080 16GB

NVIDIA16GB VRAM$999ollama, llama.cpp

NVIDIA RTX 5080 16GB profile for local AI model planning.

Vision 89Image 89Chat 80

Mac mini M4 Pro 32GB

Apple32GB unified$1,599ollama, llama.cpp

Mac mini M4 Pro 32GB profile for local AI model planning.

Mac mini M4

Choose a memory configuration. Model fit and price are calculated per variant.

2 configurations

16GB unified memory

Apple16GB unified$699ollama, llama.cpp

Mac mini M4 16GB profile for local AI model planning.

24GB unified memory

Apple24GB unified$1,099ollama, llama.cpp

Mac mini M4 24GB profile for local AI model planning.

MacBook Pro M4 Max

Choose a memory configuration. Model fit and price are calculated per variant.

2 configurations

64GB unified memory

Apple64GB unified$5,299ollama, llama.cpp

MacBook Pro M4 Max 64GB profile for local AI model planning.

128GB unified memory

Apple128GB unified$7,199ollama, llama.cpp

MacBook Pro M4 Max 128GB profile for local AI model planning.

MacBook Pro M4

Choose a memory configuration. Model fit and price are calculated per variant.

3 configurations

16GB unified memory

Apple16GB unified$1,999ollama, llama.cpp

Base M4 MacBook Pro configuration with 16GB unified memory. This is the exact hardware used for LocalAIRun's maintained Stable Diffusion 1.5 ComfyUI recipe.

24GB unified memory

Apple24GB unified$2,399ollama, llama.cpp

Base M4 MacBook Pro configuration with 24GB unified memory, grouped with the 16GB and 32GB variants while retaining its own model-fit and price profile.

32GB unified memory

Apple32GB unified$2,799ollama, llama.cpp

Base M4 MacBook Pro configuration with 32GB unified memory. It has more fit headroom than the 16GB variant but retains the same base M4 compute profile.