A curated catalogue of open source models, downloaded straight from Hugging Face and verified on arrival. Pick by memory, by use case, or let the flagship do the work.
The flagship. A reasoning model sized for laptops, quick enough to feel instant on Apple Silicon.
| Family | Sizes | Best for | Suggested memory |
|---|---|---|---|
| NVIDIA Nemotron | Nano 4B and up | Reasoning, everyday chat, the default choice | 8 GB and above |
| Llama 3 | 1B to 70B | General assistance, writing, long context | 8 GB to 64 GB |
| Qwen 3 | 0.6B to 32B | Multilingual work and coding | 8 GB to 48 GB |
| Gemma 3 | 1B to 27B | Compact quality on smaller machines | 8 GB to 32 GB |
| Phi 4 | mini to 14B | Fast responses on modest hardware | 8 GB to 24 GB |
| DeepSeek R1 Distill | 1.5B to 32B | Step-by-step reasoning and maths | 8 GB to 48 GB |
| Mistral | 7B and Small | Balanced, dependable generalists | 16 GB and above |
Guidance, not gatekeeping. The in-app catalogue reads your machine's memory and tells you what will run comfortably before you download anything.
Every model in the catalogue is open source with published weights. Downloads are resumable, so a dropped connection costs you minutes, not gigabytes.
Each downloaded file is integrity-checked against a per-file SHA-256 hash before it can ever be loaded. If a byte is wrong, the model does not run.
Pathway Lite runs models exactly as their creators published them on Hugging Face. No retraining, no fine-tuning and no modification, so what you download is what the community reviewed.
Rewrite the system prompt to shape tone, format and behaviour, per chat or per project. No vendor defaults you cannot see or change.
Turn it down for precise, repeatable answers, or up for looser, more creative output. Adjust it mid-conversation and see the difference immediately.
Pathway Lite
Llama 3
Qwen 3
Gemma 3
Phi 4
DeepSeek R1 Distill
Mistral