# Choose a Whisper model. Dimmy Help.

> Choose a Whisper model. Step-by-step help for Dimmy, the native voice dictation app for macOS, Windows, and Linux.

Source: https://dimmy.app/help/whisper-models/

[← Back to Help center](https://dimmy.app/help/)

Providers & Models

# Choose a local model

The whole on-device catalogue: nine Whisper sizes, Parakeet TDT v3, four Qwen3-ASR variants and Whistle. What each one weighs and what it runs on.

Open Settings → Voice input → Local model. On macOS the picker is grouped into Whisper, Parakeet and Qwen3-ASR, and every row names the processor it will use. On Windows it is one flat list with a VRAM dot per row. Linux does not expose a local model picker at all: the AppImage runs Whisper on the CPU.

The local model picker in Settings, with sizes and download state per row

Four engines share that one picker. If you are not sure which family you want, start from [Which model should you use?](https://dimmy.app/help/choose-your-model/). What follows is the reference: what exists, what it weighs, what it is good at.

Whistle is the newest and the smallest, added in 0.7.13: about 17 MB, and it runs on the CPU alone, so it works on a machine with no graphics card. It covers English, German, French, Spanish, Italian, Dutch and Polish. On our test machine a one-minute Italian dictation came back in under five seconds. The engine is downloaded together with the model the first time you pick it, and your custom dictionary is passed to it as keywords.

## Whisper

The broadest option: 99 languages, nine sizes, the same list on every platform.

| Model                | File                             | Size   | Good for                                                |
| -------------------- | -------------------------------- | ------ | ------------------------------------------------------- |
| Tiny                 | ggml-tiny-q8\_0.bin              | 42 MB  | Fastest, lower accuracy                                 |
| Base                 | ggml-base-q8\_0.bin              | 78 MB  | Good balance of speed and accuracy. The Whisper default |
| Small                | ggml-small-q5\_1.bin             | 181 MB | High accuracy, slower                                   |
| Medium               | ggml-medium-q5\_0.bin            | 514 MB | Very high accuracy, wants 2 GB or more of RAM           |
| Large-v3-Turbo Q5    | ggml-large-v3-turbo-q5\_0.bin    | 574 MB | Fast and accurate, all languages                        |
| Large-v3-Turbo Q8    | ggml-large-v3-turbo-q8\_0.bin    | 874 MB | Best turbo quality, all languages                       |
| Large-v3 Q5          | ggml-large-v3-q5\_0.bin          | 1.1 GB | Maximum accuracy, all languages, slow                   |
| Distil-Large-v3.5 Q8 | ggml-distil-large-v3.5-q8\_0.bin | 818 MB | English only, around 6x faster than Large-v3            |
| Distil-Large-v3.5 Q5 | ggml-distil-large-v3.5-q5\_0.bin | 538 MB | English only, compact                                   |

The q5\_0, q5\_1 and q8\_0 in the file names are quantisation levels. A lower number means a smaller file and slightly coarser weights. Dimmy only ships quantised builds, so there is no full-precision row to compare against.

## Parakeet TDT v3

One row in the picker, two very different builds underneath. On macOS it is a CoreML bundle of about 466 MB that runs on the Neural Engine. On Windows it is an ONNX bundle of about 2.5 GB that runs on the CPU. Same model, same quality, wildly different download.

* Fast, and strong on European languages.
* It takes no language argument at all, so the Language setting has no effect on it. See [Pick the dictation language](https://dimmy.app/help/language/).
* On Windows, anything longer than two minutes is split and stitched back together automatically, because the model itself breaks somewhere past five minutes of audio in one go.

## Qwen3-ASR

The newest engine, and the one that measured best on messy conversational speech and on acronyms. It needs its language named rather than left on Auto.

| Variant                               | Size   | Runs on       | Good for                                                 |
| ------------------------------------- | ------ | ------------- | -------------------------------------------------------- |
| Qwen3-ASR 0.6B                        | 971 MB | GPU or CPU    | Twice as fast, weaker on acronyms                        |
| Qwen3-ASR 1.7B                        | 2.4 GB | GPU or CPU    | Best accuracy on conversational speech. The Qwen default |
| Qwen3-ASR 0.6B · Neural Engine (int8) | 2.9 GB | Neural Engine | Leaves the GPU to the rest of the Mac. Needs macOS 15    |
| Qwen3-ASR 0.6B · Neural Engine (f32)  | 4.0 GB | Neural Engine | Full precision on the Neural Engine. Needs macOS 15      |

Each GPU variant is a pair of files, the model plus an audio projector, and both have to be on disk before the row counts as downloaded. The two Neural Engine variants only appear on Apple Silicon running macOS 15 or newer; on any other machine they are not listed rather than offered and then failing.

## What the rows tell you

* A green check means the model is already on disk. Rows you have not downloaded show their size instead.
* On macOS each row also names where it will run. A Whisper row says Neural Engine only when that model's own Core ML encoder is on disk, and GPU otherwise.
* On Windows each row can carry a coloured dot read from your graphics card: green means it fits comfortably, amber means it fits but may fail once Whisper is also resident, red means it is too large. No dot means Dimmy could not read your VRAM and will not guess.
* Nothing is ever hidden or blocked because of your hardware. The dots are advice, not a gate.

## Download flow

1. Pick a model from the list.
2. Click Download. A progress bar shows the current MB of the total, and an interrupted download resumes where it stopped rather than starting over.
3. When it finishes the model becomes active automatically.
4. Earlier models stay on disk.

info

There is no Remove button today. Models you no longer want stay in Dimmy's models folder until you delete the files yourself. If you select an engine whose files are missing, Dimmy falls back to Whisper and tells you it did, instead of failing on every dictation.

## The Neural Engine row on macOS

Whisper's encoder can move off the GPU and onto the Neural Engine, which is the difference between a long meeting slowing Dimmy down and a long meeting slowing the whole Mac down. Dimmy offers the encoder as a separate download for the selected model, around 1.2 GB for the large ones. The first transcription afterwards takes a few minutes while macOS compiles the model for your specific machine, once. The Distil models have no Core ML encoder, so no offer appears for them.

tip

Not sure which row to pick? [Which model should you use?](https://dimmy.app/help/choose-your-model/) walks through it in four questions. For how local mode works in general, see [Local mode](https://dimmy.app/help/local-mode/). The on-device LLM picker on the Output page behaves the same way, see [the on-device LLM](https://dimmy.app/help/local-llm/).

Was this helpful? 

## In this category

[ Providers & ModelsWhich model should you use?Four questions decide it: what machine you have, what must not leave it, which subscriptions you already pay for, and which language you speak. Answer them and the rest follows.](https://dimmy.app/help/choose-your-model/) [ Providers & ModelsLocal mode: transcription on your own machineFour speech engines that run on your device. No API key, no upload, no per-minute cost. Works on a plane.](https://dimmy.app/help/local-mode/) [ Providers & ModelsCloud providers comparedGroq is fastest, OpenAI is the gold standard, Deepgram leads on streaming. All optional. Bring your own key.](https://dimmy.app/help/cloud-providers/) [ Providers & ModelsAdd your API keyOpen Settings → Providers & keys, find the provider card, paste the key. It is encrypted on your machine. You can update keys at any time.](https://dimmy.app/help/api-keys/) [ Providers & ModelsOptional on-device LLMRun the LLM enhancement pass locally via llama.cpp. No cloud, no key, no per-token cost. Pick a model to match your GPU memory.](https://dimmy.app/help/local-llm/)

Last updated: 2026-05-22

On this page

1. [01 Whisper](#whisper)
2. [02 Parakeet TDT v3](#parakeet-tdt-v3)
3. [03 Qwen3-ASR](#qwen3-asr)
4. [04 What the rows tell you](#what-the-rows-tell-you)
5. [05 Download flow](#download-flow)
6. [06 The Neural Engine row on macOS](#the-neural-engine-row-on-macos)