The Open Weights
LatestModelsLeaderboardsCompanies
Subscribe
The Open Weights

The daily record of open-source AI. New model releases, leaderboards, and what's coming next — written for people who ship.

Refreshed every 12 hours

Discover

  • Latest releases
  • New today
  • Trending models

Browse

  • All models
  • Companies
  • Categories
  • Leaderboards

About

  • About
  • Editorial policy
  • Hardware estimates
  • RSS feed
  • llms.txt
  • Newsletter

© 2026 The Open Weights. An independent publication.

PrivacyTermsSMSAggregated by Claude · curated by humans.
LatestZyphraZonos3-Mini
ZyphraText → Speech

Zyphra releases Zonos3-Mini multilingual TTS

The compact open text-to-speech model supports multiple languages and voice cloning from short audio samples.

Nov 9, 2026
NotableOther
Zonos3-Mini

Zyphra has published Zonos3-Mini, a compact multilingual text-to-speech model that adds voice cloning to an increasingly crowded field of open speech systems. The model is available now on Hugging Face.

As its name suggests, Zonos3-Mini is positioned as a lightweight entry in Zyphra's Zonos line, aimed at developers who want to generate speech across languages without reaching for a large, resource-hungry model. The release lists multilingual support and voice cloning as its headline capabilities, letting users synthesize speech that mirrors a reference voice.

Why it matters

Open text-to-speech has become one of the most practical frontiers in open-weights AI, powering everything from accessibility tools to voice interfaces and localized content. A smaller model that still handles multiple languages lowers the barrier for teams that can't run heavyweight audio pipelines.

Key points from the release:

  • Multilingual synthesis in a compact "Mini" package
  • Built-in voice cloning from reference audio
  • Distributed under a custom license, so commercial users should review the terms

Zyphra has not published parameter counts or detailed benchmark figures alongside the listing, so independent evaluation of quality and latency will depend on community testing. For now, the model's arrival on Hugging Face gives developers a ready way to try multilingual cloning on their own hardware.

Sources

  • Zyphra/ZONOS3-Mini

    Hugging Face

    Visit
OlderNVIDIA Releases Cosmos 3 for World SimulationNVIDIA · Image → Video · 2 weeks agoNewerMeta AI debuts Muse Ember, a compact open modelMeta AI · Text / LLM · last week

Get the model

Hugging Face

Specs

Languagesmultilingual
Architecturemect_spk
LicenseOTHER
Downloads142.8K
Likes590

Can you run it?

Runs on just about anything.

  • BF16 (as published)

    Any modern laptop · 8 GB RAM

    1 GB
  • FP8

    Any modern laptop · 8 GB RAM

    1 GB
Your machine
  • Precision not published; assuming 16-bit (BF16) weights.

Based on Hugging Face weights metadata. Assumes an 8K context and ~1 GB runtime overhead; actual needs vary. How we estimate


Modalities

Text → Speech

The Weekly Weights

Every open release that mattered, one email a week.

0 comments

No comments yet. Be the first to weigh in.

More from Zyphra

All Zyphra releases →
Zonos 2
Zyphra/Text → Speech

Zyphra Releases Open-Source Zonos 2 TTS Model

The new text-to-speech model offers a commercially permissive alternative for developers in a field still dominated by closed-source APIs.

Jun 11, 2026

More in Text → Speech

All Text → Speech →
Haru-4B
Higurashi Ai/Any-to-AnyRuns on a laptop

Haru-4B Brings Full-Duplex Speech to Six Languages

A compact 4-billion-parameter model handles real-time, two-way voice conversation across six languages in a single system.

Nov 12, 2026
OpenBMB/Text → Speech

OpenBMB's Phonon Brings High-Quality TTS On-Device

The new text-to-speech model promises benchmark-topping voice synthesis that runs locally rather than in the cloud.

Nov 5, 2026
ServiceNow/Text → SpeechConsumer GPU

ServiceNow Debuts Apriel Voice 27B, an Open Speech Model

The 27-billion-parameter preview unifies text-to-speech, transcription, and reasoning in a single open model.

Oct 28, 2026