---
format: "aidr-story-markdown/v1"
id: "8e671fda16b4ef2f8e723ff198d4bee0366e171d5ba2a3096d223acbf9f9eed7"
canonical_url: "https://aidr.today/8e671fda?lang=en"
title: "Whistle: Speech to Text in 16.9 MB"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-10-09T05:35:51.000Z"
category: "Open Source"
topics: ["open-source"]
source_urls: ["https://cactuscompute.com/blog/whistle","https://lobste.rs/s/lpomuo/whistle_speech_text_16_9_mb"]
summary: "Whistle is an open speech recognition model that transcribes seven languages and reaches the first token in 11 ms."
---

# Whistle: Speech to Text in 16\.9 MB

> [Open the canonical story](<https://aidr.today/8e671fda?lang=en>)

**Published:** 2026-10-09T05:35:51.000Z
**Category:** Open Source
**Topics:** open\-source

## Summary

Whistle is an open speech recognition model that transcribes seven languages and reaches the first token in 11 ms\.

## Sources

- [Story source](<https://cactuscompute.com/blog/whistle>)
- [Discussion](<https://lobste.rs/s/lpomuo/whistle_speech_text_16_9_mb>)