A model is still running

A model is still running or downloading. Leaving will stop the current operation. Stay on this page to let it finish.

Skip to content
LocalAI.toolsBeta

Join the waitlist

Get an email when the new apps and model runs launch.

No account needed.

LFM3 Audio 2B Transcription

Transcribe a short English WAV clip with Liquid AI LFM3 Audio.

Browser inferencePowered by WebAI · Private · Unlimited runs
About this model

Transcribe a short English WAV clip with Liquid AI LFM3 Audio.

Created by
Liquid AI
Model released
Unknown
Source repository created
Sep 8, 2026
Model weights · Q4F16
1.7 GB

Best for

  • Transcribe short, clear English speech in the browser.

Keep in mind

  • Use mono PCM16 16 kHz WAV clips up to 30 seconds. Noise, accents and overlapping speech can cause errors.
  • This candidate implements speech-to-text. Speech synthesis, realtime sessions and mixed chat are pending.
  • The measured WebGPU profile uses GPU compute and CPU operators.
  • The Liquid AI license places conditions on commercial use and redistribution.

Processed on your deviceInference runs in this browser. Input/output sharing requires consent for each run. Basic run logs include device information and approximate connection location. Model and runtime files download from external hosts. Saved downloads may be cleared by your browser.

Sources & technical details
Original model
https://huggingface.co/LiquidAI/LFM3-Audio-2B
Release verification
Unknown
Release evidence
Exact checkpoint date pending provenance review.
Source updated
Nov 3, 2026
Browser model files
LiquidAI/LFM3-Audio-2B-ONNX
Files updated
Nov 5, 2026
Artifact revision
ac6eed2a1d6b16c785bfc1edc06b623164a22477

Input

AUDIO

Sample result

LFM3 Audio 2B Transcription
Record a clip, then run inference. This does not send live audio to the model.

Settings

Maximum decoder tokens for the transcript. Higher limits allow longer answers but use more time and memory.

Output

Your transcript or answer will appear here

Choose the input task, then add a WAV clip or text, or try a named sample.

Benchmark evidence

Saved worker evidence is separate from benchmarks run on your device. It does not guarantee speed or quality on every browser.

LFM3 Audio 2B Transcription

1 configurations tested
Passed1
Failed0
TestedNov 15, 2026
RuntimeONNX Runtime Web 1.31.0-dev.20260914-8d85527a0

Quality criteria

No explicit quality gate recorded. Review per-case evidence; full certification is pending.

PrecisionDeviceResult
Q4F16GPU · WebGPUAcceleratedPassedModerate
Test setup & sample details
Quality gate
No explicit quality gate recorded. Review the per-case quality evidence; full certification is pending.
Tested
2026-11-15T22:21:46.599Z
Environment
Chrome 157 on Apple M3 Max. Q4F16 text decoder and FP16 audio encoder. Concurrent unrelated Qwen VL matrix, native build and B2 mirror. Staged verified cache.
Runtime
ONNX Runtime Web 1.31.0-dev.20260914-8d85527a0
Artifact revision
ac6eed2a1d6b16c785bfc1edc06b623164a22477
Fixture
{
  "file": "model-samples/lfm3-audio-2b/audio-native-samples/transcription-request.json",
  "url": "https://assets.localai.tools/samples/v1/files/493b70e3171ef13e949689006ced9153fc0bcb17df6b1da46c2496c85fc75afa/transcription-request.json",
  "size": 675,
  "sha256": "493b70e3171ef13e949689006ced9153fc0bcb17df6b1da46c2496c85fc75afa",
  "sha1": "10f435066c8410ca3b83fc4330757682d4563232"
}
Rerun when
artifact revision; runtime version; fixture checksum; quality gate; backend behavior
Per-case quality & failure details
[
  {
    "precision": "q4f16",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "metric": "wordErrorRate",
      "score": 0,
      "passed": true
    },
    "runtime": {
      "engine": "ONNX Runtime Web",
      "version": "1.31.0-dev.20260914-8d85527a0",
      "device": "webgpu",
      "precision": "q4f16",
      "artifactRevision": "ac6eed2a1d6b16c785bfc1edc06b623164a22477",
      "loadMs": 3864.899999976158,
      "inferenceMs": 1719.800000011921,
      "totalMs": 5584.699999988079,
      "performanceMetric": {
        "name": "audioSecondsPerSecond",
        "value": 6.396092568859026
      },
      "grade": "fast"
    }
  }
]

LFM3 Audio 2B Spoken Questions, Transcription and Speech

12 configurations tested
Passed12
Failed0
TestedNov 15, 2026
RuntimeONNX Runtime Web 1.31.0-dev.20260914-8d85527a0

Quality criteria

No explicit quality gate recorded. Review per-case evidence; full certification is pending.

PrecisionDeviceResult
Q8F16GPU · WebGPUAcceleratedPassedFast
Q4F16GPU · WebGPUAcceleratedPassedFast
Q8F16GPU · WebGPUAcceleratedPassedFast
Q4F16GPU · WebGPUAcceleratedPassedModerate
Q8F16GPU · WebGPUAcceleratedPassedModerate
Q4F16GPU · WebGPUAcceleratedPassedModerate
Q4F16GPU · WebGPUAcceleratedPassedSlow
Q8F16GPU · WebGPUAcceleratedPassedSlow
Q8F16CPU · WASMPassedSlow
Q4F16CPU · WASMPassedSlow
Q8F16CPU · WASMPassedSlow
Q4F16CPU · WASMPassedSlow
Test setup & sample details
Quality gate
No explicit quality gate recorded. Review the per-case quality evidence; full certification is pending.
Tested
2026-11-15T22:21:46.599Z
Environment
Chrome 157 on Apple M3 Max. Q4F16 text decoder and FP16 audio encoder. Concurrent unrelated Qwen VL matrix, native build and B2 mirror. Staged verified cache.
Runtime
ONNX Runtime Web 1.31.0-dev.20260914-8d85527a0
Artifact revision
ac6eed2a1d6b16c785bfc1edc06b623164a22477
Fixture
{
  "file": "model-samples/lfm3-audio-2b-speech/audio-native-samples/transcription-request.json",
  "url": "https://assets.localai.tools/samples/v1/files/493b70e3171ef13e949689006ced9153fc0bcb17df6b1da46c2496c85fc75afa/transcription-request.json",
  "size": 675,
  "sha256": "493b70e3171ef13e949689006ced9153fc0bcb17df6b1da46c2496c85fc75afa",
  "sha1": "10f435066c8410ca3b83fc4330757682d4563232"
}
Rerun when
artifact revision; runtime version; fixture checksum; quality gate; backend behavior
Per-case quality & failure details
[
  {
    "precision": "q4f16",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "metric": "wordErrorRate",
      "score": 0,
      "passed": true
    },
    "runtime": {
      "engine": "ONNX Runtime Web",
      "version": "1.31.0-dev.20260914-8d85527a0",
      "device": "webgpu",
      "precision": "q4f16",
      "artifactRevision": "ac6eed2a1d6b16c785bfc1edc06b623164a22477",
      "loadMs": 3864.899999976158,
      "inferenceMs": 1719.800000011921,
      "totalMs": 5584.699999988079,
      "performanceMetric": {
        "name": "audioSecondsPerSecond",
        "value": 6.396092568859026
      },
      "grade": "fast"
    }
  },
  {
    "precision": "q4f16",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "metric": "wordErrorRate",
      "score": 0,
      "passed": true,
      "expected": "the library opens at nine in the morning"
    },
    "runtime": {
      "loadMs": 4140.805000007153,
      "inferenceMs": 2642.110000014305,
      "totalMs": 6782.915000021458,
      "streamMs": 2548.6299999952316,
      "durationMs": 2880,
      "firstChunkMs": 573.5949999690056,
      "precision": "q4f16",
      "device": "webgpu",
      "placement": "hybrid",
      "engine": "ONNX Runtime Web",
      "version": "1.31.0-dev.20260914-8d85527a0",
      "artifactRevision": "ac6eed2a1d6b16c785bfc1edc06b623164a22477",
      "performanceMetric": {
        "name": "wordErrorRate",
        "value": 0
      },
      "grade": "excellent"
    }
  },
  {
    "precision": "q4f16",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "metric": "exactAnswer",
      "score": 1,
      "passed": true,
      "expected": "42"
    },
    "runtime": {
      "loadMs": 4027.550000011921,
      "inferenceMs": 505.4950000047684,
      "totalMs": 4533.045000016689,
      "precision": "q4f16",
      "device": "webgpu",
      "artifactRevision": "ac6eed2a1d6b16c785bfc1edc06b623164a22477",
      "engine": "ONNX Runtime Web",
      "version": "1.31.0-dev.20260914-8d85527a0",
      "placement": "hybrid",
      "threads": 1,
      "performanceMetric": {
        "name": "exactAnswer",
        "value": 1
      },
      "grade": "excellent"
    }
  },
  {
    "precision": "q4f16",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "metric": "exactAnswer",
      "score": 1,
      "passed": true,
      "expected": "42"
    },
    "runtime": {
      "loadMs": 3714.149999976158,
      "inferenceMs": 3069.2400000095367,
      "totalMs": 6783.389999985695,
      "precision": "q4f16",
      "device": "wasm",
      "artifactRevision": "ac6eed2a1d6b16c785bfc1edc06b623164a22477",
      "engine": "ONNX Runtime Web",
      "version": "1.31.0-dev.20260914-8d85527a0",
      "placement": "wasm",
      "threads": 8,
      "performanceMetric": {
        "name": "exactAnswer",
        "value": 1
      },
      "grade": "excellent"
    }
  },
  {
    "precision": "q4f16",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "metric": "exactAnswer",
      "score": 1,
      "passed": true,
      "expected": "Paris"
    },
    "runtime": {
      "loadMs": 4027.550000011921,
      "inferenceMs": 306.5049999952316,
      "totalMs": 4334.055000007153,
      "precision": "q4f16",
      "device": "webgpu",
      "artifactRevision": "ac6eed2a1d6b16c785bfc1edc06b623164a22477",
      "engine": "ONNX Runtime Web",
      "version": "1.31.0-dev.20260914-8d85527a0",
      "placement": "hybrid",
      "threads": 1,
      "performanceMetric": {
        "name": "exactAnswer",
        "value": 1
      },
      "grade": "excellent"
    }
  },
  {
    "precision": "q4f16",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "metric": "exactAnswer",
      "score": 1,
      "passed": true,
      "expected": "Paris"
    },
    "runtime": {
      "loadMs": 3714.149999976158,
      "inferenceMs": 2605.6200000047684,
      "totalMs": 6319.7699999809265,
      "precision": "q4f16",
      "device": "wasm",
      "artifactRevision": "ac6eed2a1d6b16c785bfc1edc06b623164a22477",
      "engine": "ONNX Runtime Web",
      "version": "1.31.0-dev.20260914-8d85527a0",
      "placement": "wasm",
      "threads": 8,
      "performanceMetric": {
        "name": "exactAnswer",
        "value": 1
      },
      "grade": "excellent"
    }
  },
  {
    "precision": "q8f16",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "metric": "exactNormalizedAnswer",
      "score": 1,
      "passed": true,
      "expected": "42"
    },
    "runtime": {
      "loadMs": 4092.125,
      "inferenceMs": 489.6749999523163,
      "totalMs": 4581.799999952316,
      "precision": "q8f16",
      "device": "webgpu",
      "artifactRevision": "ac6eed2a1d6b16c785bfc1edc06b623164a22477",
      "engine": "ONNX Runtime Web",
      "version": "1.31.0-dev.20260914-8d85527a0",
      "placement": "hybrid",
      "threads": 1,
      "performanceMetric": {
        "name": "exactNormalizedAnswer",
        "value": 1
      },
      "grade": "excellent"
    }
  },
  {
    "precision": "q8f16",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "metric": "exactNormalizedAnswer",
      "score": 1,
      "passed": true,
      "expected": "Paris"
    },
    "runtime": {
      "loadMs": 4092.125,
      "inferenceMs": 304.55500000715256,
      "totalMs": 4396.680000007153,
      "precision": "q8f16",
      "device": "webgpu",
      "artifactRevision": "ac6eed2a1d6b16c785bfc1edc06b623164a22477",
      "engine": "ONNX Runtime Web",
      "version": "1.31.0-dev.20260914-8d85527a0",
      "placement": "hybrid",
      "threads": 1,
      "performanceMetric": {
        "name": "exactNormalizedAnswer",
        "value": 1
      },
      "grade": "excellent"
    }
  },
  {
    "precision": "q8f16",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "metric": "exactNormalizedAnswer",
      "score": 1,
      "passed": true,
      "expected": "And so, my fellow Americans, ask not what your country can do for you. Ask what you can do for your country."
    },
    "runtime": {
      "loadMs": 4092.125,
      "inferenceMs": 1167.8149999976158,
      "totalMs": 5259.939999997616,
      "precision": "q8f16",
      "device": "webgpu",
      "artifactRevision": "ac6eed2a1d6b16c785bfc1edc06b623164a22477",
      "engine": "ONNX Runtime Web",
      "version": "1.31.0-dev.20260914-8d85527a0",
      "placement": "hybrid",
      "threads": 1,
      "performanceMetric": {
        "name": "exactNormalizedAnswer",
        "value": 1
      },
      "grade": "excellent"
    }
  },
  {
    "precision": "q8f16",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "metric": "exactNormalizedAnswer",
      "score": 1,
      "passed": true,
      "expected": "42"
    },
    "runtime": {
      "loadMs": 3558.0099999904633,
      "inferenceMs": 2661.9300000071526,
      "totalMs": 6219.939999997616,
      "precision": "q8f16",
      "device": "wasm",
      "artifactRevision": "ac6eed2a1d6b16c785bfc1edc06b623164a22477",
      "engine": "ONNX Runtime Web",
      "version": "1.31.0-dev.20260914-8d85527a0",
      "placement": "wasm",
      "threads": 8,
      "performanceMetric": {
        "name": "exactNormalizedAnswer",
        "value": 1
      },
      "grade": "excellent"
    }
  },
  {
    "precision": "q8f16",
    "device": "wasm",
    "status": "pass",
    "quality": {
      "metric": "exactNormalizedAnswer",
      "score": 1,
      "passed": true,
      "expected": "Paris"
    },
    "runtime": {
      "loadMs": 3558.0099999904633,
      "inferenceMs": 2225.264999985695,
      "totalMs": 5783.274999976158,
      "precision": "q8f16",
      "device": "wasm",
      "artifactRevision": "ac6eed2a1d6b16c785bfc1edc06b623164a22477",
      "engine": "ONNX Runtime Web",
      "version": "1.31.0-dev.20260914-8d85527a0",
      "placement": "wasm",
      "threads": 8,
      "performanceMetric": {
        "name": "exactNormalizedAnswer",
        "value": 1
      },
      "grade": "excellent"
    }
  },
  {
    "precision": "q8f16",
    "device": "webgpu",
    "status": "pass",
    "quality": {
      "metric": "wordErrorRate",
      "score": 0,
      "passed": true,
      "expected": "the library opens at nine in the morning"
    },
    "runtime": {
      "loadMs": 3459.085000038147,
      "inferenceMs": 2761.8699999451637,
      "totalMs": 6220.954999983311,
      "streamMs": 2104.8449999690056,
      "durationMs": 2880,
      "firstChunkMs": 490.4399999976158,
      "precision": "q8f16",
      "device": "webgpu",
      "placement": "hybrid",
      "engine": "ONNX Runtime Web",
      "version": "1.31.0-dev.20260914-8d85527a0",
      "artifactRevision": "ac6eed2a1d6b16c785bfc1edc06b623164a22477",
      "performanceMetric": {
        "name": "wordErrorRate",
        "value": 0
      },
      "grade": "excellent"
    }
  }
]

Something not quite right?

Tell us what happened with Liquid LFM3. Loading, results, or the page itself—we’re here for all of it.

Attachments Optional

Up to 3 files · 50 MB each

Browser, OS, CPU/GPU, memory, screen, storage, and network details are included automatically. Model inputs, results, and browsing history are not collected.