
kapitsa
Runs Qwen3.6-35B-A3B 4-bit on Apple M5 with faster decoding and prompt reading.
Benchmarks
- On M5 Max 128GB with Qwen3.6-35B-A3B (4-bit), it decodes at 351 tok/s, against 235 tok/s for mlx-serve (+49%), and it reads prompts at 6539 tok/s, against 4622 tok/s for mlx-serve (+41%). Source
Listing
Listed: Passed all seven checks.
Security review
Not reviewed yet.
Install
To install it, follow the project's own instructions on GitHub. The engine runs with your user's permissions. macOS doesn't sandbox it, so install only engines you trust.
Something wrong or out of date? Write to help@kalsa.io, or ask us to remove this page.