Model comparison

Conv-TasNet Libri1Mix (vokra) vs Claude 4 Opus

Capability radar, value bars, and side-by-side posture.

← Back to Compare

Conv-TasNet Libri1Mix (vokra)
Not scored
value score
Claude 4 Opus
95
value score
Conv-TasNet Libri1Mix (vokra) ctx
Claude 4 Opus ctx
200K

Capability radar

Value · context · multimodal · openness · speed posture

Value head-to-head

  • Claude 4 Opus95

Attribute tape

FieldConv-TasNet Libri1Mix (vokra)Claude 4 Opus
DeveloperVokraAnthropic
Context200K
ModalitiesText, Code, Image
OpennessProprietary Model
Speed
Price$15/1M in · $75/1M out
Value score95
SummaryA Conv-TasNet speech enhancement model trained on the Libri1Mix enh_single task at 16 kHz, capable of separating and cleaning speech from noisy single-channel audio mixtures entirely in the time domain. It is derived from the Asteroid librimix recipe and closely mirrors the JorisCos/ConvTasNet_Libri1Mix_enhsingle_16k reference artifact.Anthropic frontier Opus-class model.
Open Conv-TasNet Libri1Mix (vokra)Open Claude 4 Opus