speaker-diarization-3.1 vs whisper-small

Specs, VRAM requirements and download trends — updated 10 November 2026.

pyannote · A · Open weights
Params
—
Context
—
30d
6.7 M
mit · commercial OK
openai · B · Open weights
Params
240 M
Context
—
30d
4.4 M
apache-2.0 · commercial OK

Specification comparison

Differences are highlighted; identical values are muted.

Specification speaker-diarization-3.1 whisper-small
Parameters — 240 M
Architecture — WhisperForConditionalGeneration
Context length — —
Licence mit apache-2.0
Commercial use Allowed Allowed
Languages — 99
Downloads 30d 6,732,606 4,376,031
Downloads all time 359.3 M 141.9 M
Quantizations on the Hub — SAFETENSORS
Gated Yes — accept licence No
First seen on the Hub 2023-11-16 2022-09-26

Download trend

Daily snapshots, last 90 days (13 Aug – 10 Nov)

10.0 M2.2 M
speaker-diarization-3.1 whisper-small

VRAM side by side

On RTX 4090 · 24 GB · 8K context unless noted

whisper-small · fp16 1.6 GB / 24 GB
✅ Runs comfortably

Adjacent comparisons