# Liquid AI LFM3 audio and omni evaluation: Voice Benchmark Profile

> Liquid AI reports 6.23% English ASR word error rate, 3.83 UTMOSv2 English speech quality and 73.13 on VoiceBench for LFM3-Audio-2B, and 70.09 on Daily-Omni with 46.32 on OmniBench for LFM3-Omni-2B.

- Scope: English and multilingual speech recognition, English text to speech, spoken question answering (VoiceBench), audio tool calling, audio understanding (MMAR), and for the omni model, audio-visual video understanding, image-plus-audio understanding and spoken visual question answering
- Measurement lanes: Voice experience, Spoken reasoning, Task completion
- Primary metric: Word and character error rates (%, lower is better); UTMOSv2 on a 1-5 scale and accuracy or benchmark scores (higher is better)
- Available evidence: Exact values in the Hugging Face model cards and the launch post
- Owner: Liquid AI
- Source snapshots refreshed: 2026-10-25

- Provider report retrieved: 2026-10-27

## Speech and audio

Liquid AI-reported. ASR rows are lower-is-better; the other columns are higher-is-better.

| Model | English ASR WER | Multilingual ASR WER / CER | English TTS UTMOSv2 | VoiceBench | Tool calling (internal) | MMAR |
| --- | --- | --- | --- | --- | --- | --- |
| [LFM3-Audio-2B](/models/lfm3-audio-2b) | 6.23% | 6.03% | 3.83 | 73.13 | 58.50 | 50.60 |
| [LFM3-Omni-2B](/models/lfm3-omni-2b) | 6.36% | 6.72% | 3.80 | 64.73 | 52.20 | 51.40 |

## Audio-visual understanding

Liquid AI-reported multiple-choice accuracy. Daily-Omni, WorldSense and AV-SpeakerBench use video with its soundtrack; OmniBench pairs an image with an audio clip, over all 1,142 questions; the speech column is its speech-audio subset.

| Model | Daily-Omni | WorldSense | AV-SpeakerBench | OmniBench overall | OmniBench speech |
| --- | --- | --- | --- | --- | --- |
| [LFM3-Omni-2B](/models/lfm3-omni-2b) | 70.09 | 47.10 | 46.79 | 46.32 | 48.25 |

## Spoken visual question answering

Liquid AI-reported. Image benchmarks with each question given only as speech. POPE is F1 over the adversarial, popular and random subsets.

| Model | RealWorldQA | MMStar | POPE | ERQA | EmbSpatial |
| --- | --- | --- | --- | --- | --- |
| [LFM3-Omni-2B](/models/lfm3-omni-2b) | 64.68 | 51.79 | 83.78 | 39.43 | 67.59 |

## Vision with written questions

Liquid AI-reported. OCRBench is out of 1,000; DocVQA and InfoVQA use ANLS; MMMU-Pro is the vision setting; Video-MME is without subtitles and without audio.

| Model | MMStar | MMMU (val) | MMMU-Pro | MMBench | RealWorldQA | MathVista (mini) | DocVQA | InfoVQA | OCRBench | RefCOCO | MVBench | Video-MME | RoboSpatial-Home |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| [LFM3-Omni-2B](/models/lfm3-omni-2b) | 55.67 | 42.44 | 27.28 | 73.68 | 66.54 | 62.50 | 92.47 | 64.08 | 806 | 73.55 | 58.03 | 57.70 | 49.14 |

## Interpretation limit

Liquid AI measured every result, including the competitor columns not carried here, with each model’s own chat template and greedy text decoding. English ASR averages the ESB / Open ASR suite; multilingual ASR averages FLEURS over nine languages, with CER for Chinese, Japanese and Korean. VoiceBench is the mean of 9 subsets with the three 1-5 judge scores multiplied by 20. The tool-calling score is an internal evaluation without a public task set. These results do not enter text-model or voice-agent rankings.

## Primary sources

- [Owner page](https://www.liquid.ai/blog/lfm3)
- [LFM3 launch post](https://www.liquid.ai/blog/lfm3)
- [LFM3-Audio-2B model card](https://huggingface.co/LiquidAI/LFM3-Audio-2B/blob/9d49701c372d6011d900f6abb40c6b296fb706dd/README.md)
- [LFM3-Omni-2B model card](https://huggingface.co/LiquidAI/LFM3-Omni-2B/blob/ce2458cbce4ad970f92032ed94e3d18d30b04f9c/README.md)

Canonical page: https://benchlm.ai/voice-benchmarks/liquid-lfm3-audio-omni-evaluation