Anubis
A native macOS app for benchmarking and comparing local large language models across any OpenAI-compatible backend, with real-time hardware telemetry alongside inference performance history.
Anubis is a SwiftUI app for Apple Silicon that benchmarks, compares and manages local LLMs served through Ollama, MLX, LM Studio, OpenWebUI, Docker Models and other OpenAI-compatible endpoints. It correlates real-time hardware telemetry with saved inference history, and can pull models directly from within the app.
Its measurement pipeline reports tokens per second counting all generated tokens including reasoning, separates model load time from time-to-first-token, and uses server-reported timing from backends that provide it, such as llama.cpp's timing block or LM Studio's native chat API. A Flow Builder lets a user sequence multi-model, multi-prompt benchmark runs and export share-ready report cards.