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.
LatestEdge01
Edge0Speech → Text

Audio8-ASR-Infinite brings streaming bilingual speech recognition

A new open model targets real-time transcription for Chinese and English audio.

Sep 21, 2026
NotableOther
Audio8-ASR-Infinite

A new automatic speech recognition model called Audio8-ASR-Infinite has landed on Hugging Face from Edge0, positioned as a streaming, real-time system for transcribing both Chinese and English audio. According to its model page, the focus is on low-latency, continuous recognition rather than batch transcription of pre-recorded files.

The emphasis on streaming is the key detail here. Real-time ASR is a different engineering problem from offline transcription: the model has to emit words as audio arrives, without waiting for a full utterance, which makes it useful for live captioning, voice assistants, and interactive applications.

Why it matters

Bilingual coverage of Chinese and English addresses one of the most common language pairs in production speech systems, and doing it in a single streaming model can simplify deployment.

  • Supports Chinese and English transcription
  • Designed for streaming, real-time use
  • Distributed openly through Hugging Face

As an initial release, the model arrives without published benchmarks or parameter details, so teams will want to test it against their own audio and existing baselines. Still, another open streaming ASR option is a welcome addition for developers building live voice features who prefer self-hosted models over closed APIs.

Sources

  • Edge0/Audio8-ASR-Infinite

    Hugging Face

    Visit
OlderJetBrains ships Mellum 2.1, a reasoning MoE for codeJetBrains · Code · 2 months agoNewerXiaomi's MiMo V2.6-Pro-RL Targets Agentic Multimodal WorkXiaomi · Vision-Language · 2 months ago

Get the model

Hugging Face

Specs

Languageszh, en
Size8.2 GB
PrecisionBF16
ArchitectureAudio8ASRInfiniteForConditionalGeneration
LicenseOTHER
Downloads469
Likes878

Can you run it?

Runs on a laptop — about 5.1 GB at FP8.

  • BF16 (as published)

    12 GB GPU (RTX 3060 / 4070) · 16 GB Mac

    9.2 GB
  • FP8

    8 GB GPU (RTX 4060) · 16 GB laptop or Mac

    5.1 GB
Your machine

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


Modalities

Speech → Text

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 Edge0

All Edge0 releases →
Audio8-TTS-Realistic-Air
Edge0/Text → SpeechRuns anywhere

Audio8-TTS-Realistic-Air brings zero-shot voice cloning

A new multilingual text-to-speech model promises realistic speech synthesis across languages from short voice samples.

Oct 14, 2026
Edge0-35B-A3B-preview
Edge0/Text / LLMConsumer GPU

Edge0's 35B MoE Aims for SSD-Backed Edge Inference

A preview mixture-of-experts model uses trained routing prediction to run on machines that can't hold it all in memory.

Sep 8, 2026

More in Speech → Text

All Speech → Text →
Rumble-2-Flash
Reson8/Speech → TextRuns anywhere

Reson8's Rumble-2-Flash Puts ASR in 0.5B Params

A compact open speech recognition model aims to transcribe more than 40 languages without heavyweight infrastructure.

Nov 18, 2026
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
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