Skip to content

Fact · Speech recognition

19,500times the on-chip SRAM

The memory a Whisper large model needs, about 10 GB, is roughly 19,500 times the 512 KB of internal SRAM on an ESP32-S3.

Checked against its source on October 17, 2026

What it measures
This compares the memory a Whisper large model is reported to need with the on-chip SRAM of an ESP32-S3 processor. Both figures are approximate and measure different things: the Whisper figure is for running the model, and the SRAM is the chip's own working memory.
How it is calculated
10 GB = 10,000,000 KB (1 GB = 1,000,000 KB); 10,000,000 / 512 = 19,531, rounded down to 19,500

Calculated from

Cited on

Whisper large needs about 19,500 times the ESP32-S3's SRAM | Yap