Whisper is OpenAI's family of open-source automatic speech recognition (ASR) models, released in multiple sizes and widely used for transcription and speech translation. Recent coverage compares competing open-source ASR models against it: Cactus released Whistle, a 16.9MB model scoring 4.31 WER on LibriSpeech test-clean versus 4.9 for Whisper base; Lokutor's Oído runs on an ESP32-S3 with WER 3.7/8.2 versus 6.3/15.9 for Whisper tiny.en; and a macOS local-first search tool uses Whisper for dialogue search.
Cactus Releases Whistle: A 16.9MB Speech-to-Text ModelCactus Compute released Whistle, an open-weights ASR model with 55M params (36M active) that compresses to a 16.9MB CQ2bit file and supports seven languages across 17 platforms. It scores 4.31 WER on LibriSpeech test-clean versus 4.9 for Whisper base at 145.3MB, targeting budget phones and wearables.
Local-first AI search tool for photos and video on macOSA macOS AI search tool enables semantic search across every photo and frame of video in any folder, with OCR text and Whisper dialogue search, running inference fully on-device with no accounts or cloud uploads.
Open-source Oído beats Whisper-tiny on a $5 microcontrollerThe Lokutor team released Oído, an open-source speech recognition model using a 13M-parameter int8 NVIDIA Conformer-CTC Small that runs on an ESP32-S3 with 8MB PSRAM, no GPU or NPU. It achieves LibriSpeech WER of 3.7/8.2 versus 6.3/15.9 for Whisper tiny.en.
The A.I. Daily