Fun ASR Nano 2512 HF
FunAudioLLM · released May 24, 2026 · FunAudioLLM/Fun-ASR-Nano-2512-hf
Speech to textTranscribes a recording into words
- Type
- Open weightsApache License 2.0
- Languages
- 4
- Size
- 0.8B
Context measured in tokens
Our take
Written Sep 14, 2026Fun ASR Nano is a tiny downloadable speech-to-text model built for speed and clean audio. At under a billion parameters, it runs fast enough to process an hour of audio in about seven seconds, though it covers only four languages and must be self-hosted.
Pick this when you need local transcription with minimal hardware, especially for clean read-aloud or audiobook work where it beats most alternatives. It also suits financial-call transcription at speed. Skip it if you need hosted inference, broad language support, or reliable results on podcasts, video or heavily accented speech.
The case for it
- Processes audio at 497 times real time — an hour in roughly seven seconds on reference hardware.
- Strong on clean read-aloud audio with a word error rate below the median across 85 models.
- Surprisingly capable on recorded meetings with crosstalk, again better than most despite its tiny size.
- Apache 2.0 licence allows commercial use, fine-tuning and redistribution.
The case against it
- Worse than most on podcasts and video audio, and poor on heavily accented speech.
- No commercial hosting available — you must run it yourself.
- Only four languages covered: Chinese, English, Japanese and Cantonese.
How good is it?
An open transcription model that turns speech into text, though accented recordings come out less cleanly than with most models.
- transcribing speakers with a range of accentsAccented speech · 62nd of 76
TranscriptionTurning speech into text2.5 of 5Open ASR WER · 52nd of 76
94.3%
Misses roughly one word in 17, averaged over nine English test sets.
497×48th of 74
an hour of audio in 7 seconds, on the board's own hardware. Your machine will differ.
4
Listed on the model card. The accuracy above is English only.
Percentage of words wrong on each set, lower better. Bars are scaled to this model's own worst case; the placing beneath each rate is against every model measured on that set.
Which languages ↓ ↑
Chinese · English · Japanese · Cantonese
The figures above come from the Open ASR Leaderboard, an independent public test that runs every model on the same recordings. It is the only measurement of transcription quality we know of, so there are no other scores to show.
Each of these is the same transcription job on a different kind of recording, so together they say where it holds up and where it slips — not how closely it follows an instruction.
Every published score for this model9 scoresEvery figure we hold, from 9 boards, with who ran it and a link to the source — including the boards no rating above is built on.
Can you run it yourself?
Comfortable fit
GeForce RTX 4090 · 24 GB
Room to spare. 20.9 GB spare means a 10% error in the size would not change the answer.
Apple M1 Pro (16-core GPU) · 32 GB
Room to spare. 22.1 GB spare means a 10% error in the size would not change the answer.
Apple M2 (8-core GPU, 8GB unified) · 8 GB
Room to spare. 4.1 GB spare means a 10% error in the size would not change the answer.
These cards answer whether Fun ASR Nano 2512 HF loads, not how fast it transcribes. Throughput figures for a transcription model come from its text decoder, so treat this as a fit answer rather than a speed one.
Memory use by level
Against a 24 GB card.
What is quantisation? →This model on every device we track71 devicesThe Q4 build most people download, on each device: what the weights come to, how much context the memory leaves, and whether it runs. Smallest device that runs it first. This is a fit answer: whether it loads, not how fast it transcribes.
When we formed this view
Recent changes
What moved
first indexed by our pipelineEach date is the day we first saw the change, or the day the maker announced it.
What we do not know about this model yet
- We hold no measured file for it, so all 3 sizes on this page are calculated from the parameter count.
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
Licence and identifiers
What the licence allowsApache License 2.0, what it allows commercially, and the identifiers you need to pull this model — its Hugging Face repo, our slug and a machine-readable card.
Licence
Apache License 2.0
Fully permissive: commercial use, redistribution, and derivatives allowed. Requires attribution and a copy of the license. Includes an express patent grant.
Identifiers
- Hugging Face
- FunAudioLLM/Fun-ASR-Nano-2512-hf
- Architecture
- Dense
- Takes in, gives back
- Audio in, text out
- Catalogue slug
- funaudiollm-fun-asr-nano-2512-hf