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
/
XHToken
XHToken/Spark-X2.6-4B
Hugging Face
4.1B params
Spark2_5ForCausalLM
text-generation
apache-2.0
1M trained context
7.7 GiB repo
RTX 3060 12GB
✓
Fits
Q8_0
53.1 tok/s
RTX 4060 Ti 16GB
✓
Fits
Q8_0
43.9 tok/s
RTX 4090 24GB
✓
Fits
BF16
83.7 tok/s
RTX 5090 32GB
✓
Fits
BF16
138 tok/s
2x RTX 3090 (48GB)
✓
Fits
BF16
77.4 tok/s
Radeon RX 7900 XTX 24GB
✓
Fits
Q8_0
68.2 tok/s
Apple M4 32GB
✓
Fits
Q8_0
16.0 tok/s
Apple M4 Pro 48GB
✓
Fits
Q8_0
32.7 tok/s
Apple M4 Max 128GB
✓
Fits
Q8_0
56.1 tok/s
Apple M3 Ultra 512GB
✓
Fits
Q8_0
74.0 tok/s
NVIDIA DGX Spark (GB10) 128GB
✓
Fits
Q8_0
42.0 tok/s
AMD Ryzen AI Max+ 395 128GB
✓
Fits
Q8_0
23.8 tok/s
H100 80GB SXM
✓
Fits
BF16
227 tok/s
NVIDIA DGX/HGX H100 8x80GB (640GB)
✓
Fits
BF16
227 tok/s
Apple iPhone 17 Pro / Pro Max (A19 Pro) 12GB
✓
Fits
Q8_0
9.40 tok/s
Samsung Galaxy S25 Ultra (Snapdragon 8 Elite) 12GB
✓
Fits
Q8_0
7.80 tok/s
Raspberry Pi 5 16GB
✓
Fits
Q3_K_L
3.10 tok/s
NVIDIA Jetson AGX Thor 128GB
✓
Fits
Q8_0
24.1 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