# homi model sizes and sources

> homi is egma-ai's open audio decision model, based on Qwen3-Omni-30B-A3B. It returns option probabilities from audio without a transcript.

Updated 2026-11-04 (Australia/Adelaide)

By egma-ai · Hugging Face weights upload: 2026-10-27 · OPEN

An open decision model that listens instead of reading: give it an audio clip and a typed question (how does the speaker feel? does the pitch rise?) and it returns a probability for every option, without a transcript. It is Qwen3-Omni-30B-A3B with frozen weights and a decision head in place of the text output. The date is when the weights went up on Hugging Face.

## Facts

- Base: Qwen3-Omni-30B-A3B-Instruct, frozen, with a decision head and 17M parameters of LoRA
- Licence: Apache-2.0
- Inputs: Audio clip plus Noul, Choice or Score questions

## Claims and sources

- **VENDOR-REPORTED**: 73.9 on VocalGrad against 63.0 for Gemini 3.8 Flash, and 83.1 on MMSU against Gemini's 85.7. Cost per 10,000 MMSU questions: an estimated $0.28 of batched GPU time on one H200 at $4.54/hour, against $14.80 in Gemini API charges at promotional [pricing](<https://decisionmodels.ai/pricing>) [source](<https://github.com/egma-ai/homi>)

## Links

- [Source: GitHub](<https://github.com/egma-ai/homi>)
- [Source: Hugging Face](<https://huggingface.co/egma-ai/homi>)
- [Source: Show HN](<https://news.ycombinator.com/item?id=50276913>)

## Related projects

No related projects are listed in the registry.