Fun-ASR-MLT-Nano-2512
Fun-ASR-MLT-Nano-2512 is an 800M-parameter speech-to-text model from FunAudioLLM, released under the Apache-2.0 license. The reviewed source lists 31 languages (31 explicitly enumerated in our record), and the model is designed for batch transcription, with live use only through a runtime or chunking wrapper. It is a specialist rather than a general pick: its best placement in the common setups we test is #8 of 26, for Japanese live transcription. Use the matcher with your own language and hardware to see where it fits.
Compact 800M multilingual Fun-ASR checkpoint covering 31 listed languages with hotword and ITN controls. The current model card still lists timestamp return and speaker diarization as TODO items; streaming is treated as a toolkit/runtime workflow rather than a proven native streaming architecture.
Key facts
| Provider | FunAudioLLM |
|---|---|
| Model family | Fun-ASR |
| Parameters | 800M |
| License | Apache-2.0 · Permissive — license text ↗ |
| Language coverage | 31 listed languages · 31 enumerated in our record |
| Streaming | Streaming via a runtime or chunking wrapper |
| Timestamps | Not documented |
| Speaker diarization | No — needs a separate component |
| Speech translation | No |
| Documented runtimes | FunASR — how to run it |
| Primary source | huggingface.co/FunAudioLLM/Fun-ASR-MLT-Nano-2512 ↗ |
| Last reviewed |
Hardware and workload fit
These are Speech Model Match's conservative 1–5 editorial ratings, relative to the other 19 reviewed models. They come from model size, official memory guidance and documented runtimes — not from a universal benchmark. How the ratings work.
| Dimension | Rating | Scale |
|---|---|---|
| CPU-only fit | 2/5 | |
| Apple Silicon fit | 2/5 | |
| NVIDIA GPU fit | 5/5 | |
| Speed | 4/5 | |
| Accuracy (relative) | 4/5 | |
| Low memory use | 3/5 | |
| Long recordings | 4/5 | |
| Live / low latency | 3/5 | |
| Subtitle work | 2/5 | |
| Edge / low-power devices | 2/5 |
Where it is strong
- NVIDIA GPU fit: rated 5/5
- Speed: rated 4/5
- Accuracy (relative): rated 4/5
- Long recordings: rated 4/5
What to check first
- CPU-only fit: rated 2/5
- Apple Silicon fit: rated 2/5
- Subtitle work: rated 2/5
- Edge / low-power devices: rated 2/5
- No native speaker diarization — pair it with a separate diarization pipeline if you need speaker labels
- Timestamp support is limited or runtime-dependent — test it if you need subtitles
How it ranks in common setups
Ranks come from the same deterministic matcher that runs on the homepage, applied to all 26 reviewed models. They are decision guidance, not accuracy claims.
| Setup | Rank | Fit score | Top match |
|---|---|---|---|
| Long recordings on an 8–12 GB NVIDIA GPU, balanced | #9 of 26 | 75/100 · Strong fit | Whisper turbo |
| Live captions or voice input (native streaming + timestamps, speed first) | #15 of 26 | 73/100 · Strong fit | Nemotron 3.5 ASR Streaming 0.6B |
| Subtitles with timing, accuracy first | #15 of 26 | 68/100 · Possible fit | Whisper medium |
| Meetings with speaker labels (diarization + timestamps) | #14 of 26 | 72/100 · Strong fit | VibeVoice-ASR 7B |
| CPU-only machine, low memory use | #18 of 26 | 58/100 · Possible fit | VibeVoice-ASR-BitNet |
| Apple Silicon with 8–16 GB unified memory, long recordings | #20 of 26 | 61/100 · Possible fit | Whisper small |
| Japanese live transcription (native streaming, speed first) | #8 of 26 | 73/100 · Strong fit | Nemotron 3.5 ASR Streaming 0.6B |
Supported languages
- Arabic
ar - Bulgarian
bg - Cantonese
yue - Croatian
hr - Czech
cs - Danish
da - Dutch
nl - English
en - Estonian
et - Filipino
fil - Finnish
fi - Greek
el - Hindi
hi - Hungarian
hu - Indonesian
id - Irish
ga - Japanese
ja - Korean
ko - Latvian
lv - Lithuanian
lt - Malay
ms - Maltese
mt - Mandarin Chinese
zh - Polish
pl - Portuguese
pt - Romanian
ro - Slovak
sk - Slovenian
sl - Swedish
sv - Thai
th - Vietnamese
vi
Compare coverage across every reviewed model in the language support matrix.
Community activity on Hugging Face
- Downloads, last 30 days
- 9,896
- Likes
- 72
- Repository last updated
- 5 Aug 2026
- Published on the Hub
- 15 Dec 2025
Figures from the Hugging Face Hub API, 11 Nov 2026. Source: huggingface.co/FunAudioLLM/Fun-ASR-MLT-Nano-2512 ↗. Downloads count file downloads, not active users, and do not affect the matcher's scores. Compare all models.
How to run Fun-ASR-MLT-Nano-2512 locally
Fun-ASR models run through Alibaba's FunASR toolkit. Runtimes below are the ones the official model card or repository documents for this model (checked November 2, 2026).
- FunASR ↗Framework · AutoModel API
Install command shown in the official huggingface.co page ↗:
python -m pip install -U "funasr>=1.4.1"
Official hardware guidance: The model card's example sets the device to a CUDA GPU or the CPU; no memory figure is published. Source ↗
Whatever runtime you choose, measure speed and accuracy on your own audio with the benchmark tool; results differ a lot between runtimes, precision and hardware.
Before you deploy Fun-ASR-MLT-Nano-2512
A printable checklist built from the facts on this page. Tick it off before putting the model into production.
- Language. Confirm your language is in the 31-language list on this page, and test any accents or code-switching you expect.
- License. Apache-2.0: Give recipients a copy of the license, keep copyright and attribution notices, carry forward any NOTICE file and mark files you changed. The patent license ends if you start patent litigation over the work. License guide.
- Install. Start from the official command:
python -m pip install -U "funasr>=1.4.1"(full run notes). - Audio input. Convert a sample of your real recordings to the format the runtime expects before benchmarking.
- Hardware. The model card's example sets the device to a CUDA GPU or the CPU; no memory figure is published.
- Features you rely on. For Fun-ASR-MLT-Nano-2512, streaming only through a runtime or chunking wrapper; timestamps are not documented; speaker labels need a separate diarization component. Test each one you need with real audio.
- Measure. Run 10–20 representative clips, score them with the WER calculator and record speed in the benchmark workspace before you commit.
Alternatives and related models
Similar fit profile: Qwen3-ASR 0.6B (Qwen, 0.6B), Parakeet TDT 0.6B v3 (NVIDIA, 600M), Distil-Whisper large-v3 (Hugging Face / Distil-Whisper, large-v3 distilled). Similarity is computed from the ten fit ratings plus streaming and diarization support.
| Fact | Fun-ASR-MLT-Nano-2512 | Qwen3-ASR 0.6B | Parakeet TDT 0.6B v3 |
|---|---|---|---|
| Provider · size | FunAudioLLM · 800M | Qwen · 0.6B | NVIDIA · 600M |
| License | Apache-2.0 Permissive | Apache-2.0 Permissive | CC BY 4.0 Attribution required |
| Languages | 31 listed languages | 30 languages + 22 Chinese dialects | 25 European languages |
| Streaming | Via runtime | Via runtime | Via runtime |
| Timestamps | Not documented | Separate aligner | Yes |
| Speaker diarization | Separate pipeline | Separate pipeline | Separate pipeline |
| CPU · Apple · NVIDIA fit | 2/5 · 2/5 · 5/5 | 2/5 · 3/5 · 5/5 | 2/5 · 3/5 · 5/5 |
| Live · long recordings | 3/5 · 4/5 | 4/5 · 4/5 | 2/5 · 5/5 |
| Main runtime | FunASR | qwen-asr package | NVIDIA NeMo |
Head-to-head pages: Qwen3-ASR 0.6B vs Fun-ASR-MLT-Nano-2512
Compare Fun-ASR-MLT-Nano-2512 side by side
Guides that cover Fun-ASR-MLT-Nano-2512
- Best local speech-to-text models in 2026
- Best local speech-to-text models for multilingual use
- Best local Whisper alternatives in 2026
Check Fun-ASR-MLT-Nano-2512 for your setup
Pick your language, hardware and workload to see this model's fit score, the reasons behind it and where it ranks against all reviewed models. Everything runs in your browser.
Common questions
Can I use Fun-ASR-MLT-Nano-2512 commercially?
Permissive license; commercial use is allowed under the listed terms. The listed license is Apache-2.0; read the current license text ↗ before shipping a product, because terms can change. The license guide compares the obligations of all 26 reviewed models.
Does Fun-ASR-MLT-Nano-2512 support real-time streaming?
Not natively. Live use relies on a runtime or chunking wrapper around a batch model, which adds latency and can split words at chunk borders. If live latency is critical, compare it with a native streaming model such as Nemotron 3.5 ASR Streaming 0.6B.
Which languages does Fun-ASR-MLT-Nano-2512 transcribe?
31 listed languages. The full list of 31 explicitly enumerated languages is shown above.
Can Fun-ASR-MLT-Nano-2512 tell speakers apart?
Not natively. Combine it with a dedicated diarization component and evaluate that component separately; the meetings guide covers this workflow.
Primary sources
Reviewed against primary sources on September 16, 2026. Model behavior, runtime support, language coverage and licenses can change; re-check the linked primary source before production use. See the full source directory.