Skip to content
modelread
Ask
Browse
Devices
Connect
Docs
Methodology
Search
⌘K
My hardware
All
Loading devices…
System RAM / unified memory override
GB
Remembered on this device. Multi-GPU: search “2x”, “8x”, “node”.
Ask
Browse
Devices
Connect
Docs
Methodology
On your hardware
Search models, devices and pages
Type a model or device name.
Pages
Ask
what should I run?
Browse
best models per device
Devices
all hardware
Connect your agent
MCP · API
How the numbers work
calibration
Models
/
ibm-granite
ibm-granite/granite-4.2-8b
Hugging Face
8.8B params
GraniteForCausalLM
text-generation
apache-2.0
128K trained context
16.4 GiB repo
RTX 3060 12GB
✓
Fits
Q6_K
32.0 tok/s
RTX 4060 Ti 16GB
✓
Fits
Q8_0
22.1 tok/s
RTX 4090 24GB
✓
Fits
Q8_0
72.9 tok/s
RTX 5090 32GB
✓
Fits
Q8_0
121 tok/s
2x RTX 3090 (48GB)
✓
Fits
Q8_0
67.0 tok/s
Radeon RX 7900 XTX 24GB
✓
Fits
Q8_0
40.5 tok/s
Apple M4 32GB
✓
Fits
MLX-8BIT
8.30 tok/s
Apple M4 Pro 48GB
✓
Fits
MLX-8BIT
17.9 tok/s
Apple M4 Max 128GB
✓
Fits
MLX-8BIT
33.0 tok/s
Apple M3 Ultra 512GB
✓
Fits
MLX-8BIT
46.4 tok/s
NVIDIA DGX Spark (GB10) 128GB
✓
Fits
Q8_0
21.1 tok/s
AMD Ryzen AI Max+ 395 128GB
✓
Fits
Q8_0
12.4 tok/s
H100 80GB SXM
✓
Fits
BF16
134 tok/s
NVIDIA DGX/HGX H100 8x80GB (640GB)
✓
Fits
BF16
134 tok/s
Apple iPhone 17 Pro / Pro Max (A19 Pro) 12GB
✓
Fits
Q4_K_M
6.80 tok/s
Samsung Galaxy S25 Ultra (Snapdragon 8 Elite) 12GB
✓
Fits
Q3_K_L
5.80 tok/s
Raspberry Pi 5 16GB
✓
Fits
Q8_0
0.80 tok/s
NVIDIA Jetson AGX Thor 128GB
✓
Fits
Q8_0
10.5 tok/s
My hardware
All
Loading devices…
System RAM / unified memory override
GB
Remembered on this device. Multi-GPU: search “2x”, “8x”, “node”.
Context
4K
8K
16K
32K
64K
128K
256K
1M
KV cache
f16
q8_0
q4_0