File	Size	SHA-256
`Qwen3-Omni-30B-A3B-Captioner-Q4_K_M.gguf`	17.7 GiB	`f6b443caefee34c82edc51278bbd410ad58146baf6fca53d09035322300034cf`
`mmproj-Qwen3-Omni-30B-A3B-Captioner-BF16.gguf`	2.1 GiB	`81d4d92662eb7f387f09b0a6679e945bf61dd0f2ef5ab8d36be6cc9580bb6252`

Validation

The Q4_K_M model and BF16 multimodal projector were tested together with llama-mtmd-cli on an audio sample. The audio encoder processed all five chunks and generated a caption.

Usage

Use a recent llama.cpp build with audio multimodal support:

bash

llama-mtmd-cli \
  -m Qwen3-Omni-30B-A3B-Captioner-Q4_K_M.gguf \
  --mmproj mmproj-Qwen3-Omni-30B-A3B-Captioner-BF16.gguf \
  --audio sample.mp3 \
  -p "Describe this audio."

Audio support in llama.cpp remains experimental. Captioner is designed for one audio input and text output.

presencesw/Qwen3-Omni-30B-A3B-Captioner-GGUF

Qwen3-Omni-30B-A3B-Captioner-GGUF

Files

Validation

Usage

presencesw/
Qwen3-Omni-30B-A3B-Captioner-GGUF