# Mini-Omni

Towards Open-source GPT-4o with Vision, Speech and Duplex Capabilities。

Repository: https://github.com/gpt-omni/mini-omni2
Canonical: https://ross.abutalabs.com/products/mini-omni
Homepage: https://arxiv.org/abs/2410.11190
Language: Python
License: MIT
License Family: permissive
Last push: 2025-01-16T14:20:25+00:00
Link (homepage): https://arxiv.org/abs/2410.11190
Link (site_page): https://info.arxiv.org/about/donate.html
Link (site_page): https://info.arxiv.org/labs/index.html
Link (site_page): https://info.arxiv.org/about/ourmembers.html
Link (site_page): https://info.arxiv.org/about

## Health v2 (maintenance only)
Score: 22/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 1, release rhythm 35, longevity 49
- inputs: {"age_days": 688, "days_push": 594, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1922, forks 209 (observed 2026-08-28T04:05:55.203674+00:00)

## What it is
Mini-Omni is an open-source multimodal large language model that performs real-time end-to-end speech-to-speech conversation with streaming audio output, without requiring separate ASR or TTS models. Mini-Omni2 extends it with vision and audio inputs and duplex interaction, aiming to be an open-source approximation of GPT-4o.

## Use cases
- build a real-time voice chat assistant
- run speech-to-speech conversation without ASR/TTS pipelines
- stream audio responses from an LLM while it generates text
- add vision and audio understanding to a conversational model
- research multimodal speech language models
- batch audio-to-text and audio-to-audio inference

## When to choose
- you need an open-source, self-hosted GPT-4o-like voice assistant
- you want low-latency streaming speech output from an LLM
- you are researching end-to-end speech language models
- English-only voice interaction is acceptable

## When to avoid
- you need multilingual speech support (English only)
- you need production-grade, polished voice assistant infrastructure
- you lack GPU hardware for local inference
- you need high-accuracy ASR or TTS as standalone components

## Facets
- artifact type: library
- maturity: experimental
- function: speech-recognition, tts, llm-inference, machine-learning, chatbot
- domain: artificial-intelligence, large-language-models, speech-processing
- platform: python, cross-platform
- tags: speech-to-speech, streaming-audio, multimodal, voice-assistant, real-time-conversation, gpt-4o-alternative, natural-language-processing, gpu

## Member repositories
- gpt-omni/mini-omni2 (main) score 22
- gpt-omni/mini-omni (mirror) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:55.203674+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:34:07.047915+00:00, confidence not recorded.
  - readme: https://github.com/gpt-omni/mini-omni2 (fetched 2026-08-28T04:05:55.203674+00:00, sha c1edf62b32fe)
  - homepage: https://arxiv.org/abs/2410.11190 (fetched 2026-08-29T09:27:45.826954+00:00, sha bf41c7b1a643)
  - site_page: https://info.arxiv.org/about/donate.html (fetched 2026-08-29T09:27:45.836023+00:00, sha cca9c3a11c56)
  - site_page: https://info.arxiv.org/about/ourmembers.html (fetched 2026-08-29T09:27:45.839267+00:00, sha 47cbc55ff1de)
  - site_page: https://info.arxiv.org/about (fetched 2026-08-29T09:27:45.841148+00:00, sha a1f16f915a9a)
  - site_page: https://info.arxiv.org/labs/index.html (fetched 2026-08-29T09:27:45.837730+00:00, sha b14a8d05a0ec)
- Data as of 2026-08-30T08:39:29.467469+00:00.