Local Whisper AI Models

Optimized ONNX models for high-performance Speech-to-Text on your local hardware.

Whisper Tiny (Multilingual)
Tiny 150.0 MB
language

Most stable multilingual model. ~150MB.

memory
VRAM 1GB
developer_board
CPU Intel Core i3 / Apple M1
Privacy Validated verified_user
Whisper Tiny (English)
Tiny 150.0 MB
bolt

Fastest model, lower accuracy. ~150MB.

memory
VRAM 1GB
developer_board
CPU Intel Core i3 / Apple M1 / Snapdragon 8cx
Privacy Validated verified_user
Whisper Base (English)
Small 290.0 MB
balance

Excellent balance of speed and accuracy. ~290MB.

memory
VRAM 2GB
developer_board
CPU Intel Core i5 / Apple M1 / Snapdragon X Plus
Privacy Validated verified_user
Whisper Small (Multilingual)
Small 950.0 MB
mic

Better accuracy, works on older hardware. ~1GB.

memory
VRAM 2GB
developer_board
CPU Intel Core i5 / Apple M2 / Snapdragon X Elite
Privacy Validated verified_user
Whisper Medium (English)
Medium 3.0 GB
auto_awesome

Near state-of-the-art accuracy for English. ~3GB.

memory
VRAM 4GB
developer_board
CPU Intel Core i7 / Apple M2 / Snapdragon X Elite
Privacy Validated verified_user
Whisper Medium (Multilingual)
Medium 3.0 GB
translate

High‑accuracy multilingual model. ~3GB.

memory
VRAM 4GB
developer_board
CPU Intel Core i7 / Apple M2 / Snapdragon X Elite
Privacy Validated verified_user
Whisper Large v2 (Multilingual)
Large 3.2 GB
star

High‑accuracy multilingual model. ~3GB.

memory
VRAM 8GB
developer_board
CPU Intel Core i9 / Apple M3 / Snapdragon X Elite
Privacy Validated verified_user
Whisper Large v3 Turbo (Multilingual)
Large 3.2 GB
speed

Fastest high-accuracy multilingual model. ~3.2GB.

memory
VRAM 6GB
developer_board
CPU Intel Core i9 / Apple M3 / Snapdragon X Elite
Privacy Validated verified_user
Whisper Large v3 (Multilingual)
Large 3.2 GB
star

State-of-the-art multilingual model. ~3.2GB.

memory
VRAM 8GB
developer_board
CPU Intel Core i9 / Apple M3 / Snapdragon X Elite
Privacy Validated verified_user