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
/
thinkingmachines
thinkingmachines/Inkling-2-Small
Hugging Face
593B params · MoE 384 experts, top-8
Inkling2ForConditionalGeneration
image-text-to-text
apache-2.0
366.5 GiB repo
RTX 3060 12GB
✓
Fits
BF16
66.6 tok/s
RTX 4060 Ti 16GB
✕
Won't fit
BF16
RTX 4090 24GB
✕
Won't fit
BF16
RTX 5090 32GB
✕
Won't fit
BF16
2x RTX 3090 (48GB)
✕
Won't fit
BF16
Radeon RX 7900 XTX 24GB
✕
Won't fit
BF16
Apple M4 32GB
✕
Won't fit
BF16
Apple M4 Pro 48GB
✕
Won't fit
BF16
Apple M4 Max 128GB
✕
Won't fit
BF16
Apple M3 Ultra 512GB
✓
Fits
Q4_K_S
39.1 tok/s
NVIDIA DGX Spark (GB10) 128GB
✕
Won't fit
BF16
AMD Ryzen AI Max+ 395 128GB
✕
Won't fit
BF16
H100 80GB SXM
✓
Fits
BF16
247 tok/s
NVIDIA DGX/HGX H100 8x80GB (640GB)
✓
Fits
Q8_0
97.4 tok/s
Apple iPhone 17 Pro / Pro Max (A19 Pro) 12GB
✕
Won't fit
BF16
Samsung Galaxy S25 Ultra (Snapdragon 8 Elite) 12GB
✕
Won't fit
BF16
Raspberry Pi 5 16GB
✕
Won't fit
Q8_0
NVIDIA Jetson AGX Thor 128GB
✕
Won't fit
BF16
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