Models & hardware

Every model runs on your Mac. Pick one in Settings › Models; you can install several and switch any time.

Requirements

• Apple silicon Mac (M1 or later), macOS 14 or later.
• Memory: each model needs roughly the amount listed below while loaded. We expect 8 GB Macs to work, but have only measured on a 36 GB Mac so far.
• Disk space for each model you install (sizes below).

Models

Parakeet models detect which of their languages you're speaking. Whisper models detect any of 99 languages, or you can set one in Settings › Models for better accuracy.

Measured performance

Run on Mac Studio (M4 Max, 36 GB), Version 15.6 (Build 24G84), 2026-10-01, using the same code that ships in the app.

How this was measured — and its limits

• Accuracy: 40 utterances (1,082 words, 4 speakers) from LibriSpeech test-clean — clearly read audiobook speech. Real dictation is messier, so expect higher error rates. With this few words, a difference of a few errors between models isn't meaningful.
• Normalization: lowercase, punctuation removed; numbers aren't converted, so “10” vs “ten” counts as an error.
• Time to text: 8 short dictation-style clips made with macOS text-to-speech, timed from the end of audio to the transcript. Cleanup adds under a millisecond.
• First load happens once after installing a model, while macOS compiles it for your chip. Later launches are faster.
• Only one Mac was measured. Base M1 Macs will be noticeably slower; we'll publish more results as we test more hardware.

The benchmark tool and sample list are in the Sayburst source repository (bench/).

Sayburst Writer

Sayburst Writer is the optional AI model that polishes your dictation: it resolves self-corrections, lays out emails and bullets spoken lists. It's Qwen3 4B Instruct (2507), quantized to 4 bits (Q4_K_M, 2.5 GB), and runs in Sayburst on your Mac's GPU with llama.cpp.

• Download: Settings › Writing › Sayburst Writer, or during setup. Verified against its published SHA-256 checksum, like the speech models.
• Memory: about 3 GB while loaded. On Macs with 16 GB or more it stays loaded for instant polishing; with less memory it unloads after 15 minutes idle.
• Speed: about 0.3–1 second per dictation on an M4 Max in our testing; slower on base M1 and 8 GB Macs. If it's still loading when you finish talking, your dictation is typed cleaned but unpolished rather than making you wait.
• Safety: if a rewrite contains a number you didn't say, or is far longer than what you said, Sayburst types the cleaned transcript instead.

Credits

• Parakeet TDT 0.6B v3 by NVIDIA, CC BY 4.0. Converted to Core ML by FluidInference. Source · CC BY 4.0
• Parakeet Ultra by moondream, post-trained from NVIDIA Parakeet TDT 0.6B v3, CC BY 4.0. Converted to Core ML by FluidInference. Source · CC BY 4.0
• Parakeet TDT 0.6B v2 by NVIDIA, CC BY 4.0. Converted to Core ML by FluidInference. Source · CC BY 4.0
• Whisper by OpenAI, MIT. ggml conversion by the whisper.cpp project, MIT. Source · MIT
• Whisper by OpenAI, MIT. ggml conversion by the whisper.cpp project, MIT. Source · MIT
• Whisper by OpenAI, MIT. ggml conversion by the whisper.cpp project, MIT. Source · MIT
• Qwen3 4B Instruct (2507) by the Qwen team, Alibaba Cloud; GGUF conversion by Unsloth. Source · Apache 2.0
• llama.cpp by the ggml authors. Source · MIT