SAVA

Released
2026-11-18
Last refreshed
2026-11-19
Status
Researched today
Open sourceCommercial use: permitted

SAVA is a released general LLM work model with open-source; evaluate it while provider pricing coverage matures.

Inworld releases · 3 in the last 12 months · this family litChangelog →
Specifications
Family
SAVA
Released
2026-11-18
Parameters
25.5M
Specialization
voice-activity-detection
Openness
Open source
License
Apache 2.0OSI-approvedPermitted
Weights
Available
Code
Unknown
Created by

Inworld develops conversational AI and speech models.

Website
Pricing

No tracked provider pricing is available yet.

About

SAVA is Inworld’s streaming speech-activity and end-of-turn detector. Its shared network uses acoustic and sentence context to distinguish speech from other sounds and estimate whether a speaker has finished. Code and weights are published under Apache 2.0.

Providers

No active provider routes are tracked for this model yet.

Capabilities

Audio

Benchmark peer barsfor Coding

No task-mapped benchmark peers are available for this model yet.

Migration checks

No linked migration route is available for this model yet.