Models & Hardware
Hardware Profiles & Local Privacy
Choose the optimal model for your system memory while maintaining zero network telemetry.
Hardware Profiles
- 8 GB RAM (Ultralight Models (1.5B – 3B)): Qwen 2.5 Coder 1.5B, Llama 3.2 3B — Ideal for laptops without dedicated GPU. Ultra-low latency and battery-friendly.
- 16 GB RAM (Balanced High-Precision Models (7B – 8B)): Qwen 2.5 Coder 7B, DeepSeek Coder 6.7B — The recommended sweet spot for professional coding with deep semantic reasoning.
- 32 GB+ RAM (Advanced & MoE Models (14B – 32B)): Qwen 2.5 Coder 14B / 32B, Mixtral 8x7B — Handles complex multi-file refactorings and extensive token context windows with ease.