Our models
Models
Open, on-device models built for reliable tool use, native Mac automation, and private local agents.
Raptor 0.6
A 4B dense agent model built on Spark-X2.5-4B, fine-tuned on the Osaurus tool surface, and quantized to 3.7 GB — reasons by default, decodes at ~105 tokens/s on an M5 Max, and stays resident on the 8–16 GB Macs most people own.
4.1B
Dense parameters
3.7 GB
Bundle size
105 tok/s
Decode, M5 Max
Raptor 0.5
An 8B mixture-of-experts that activates only ~1B parameters per token, runs the full Osaurus tool surface in 6.3 GB, and stays resident on the 8–16 GB Macs most people own.
7.9B
Total parameters
~1B
Active per token
6.3 GB
Bundle size
AppleScript-8B & AppleScript-16B-A4B
Two open models that turn plain-English requests into working macOS automation — on-device, tool-calling, and quantized for Apple Silicon.
8B & 16B
Model sizes
8–16 GB
RAM required
100% local
Private by default