GLM ASR Nano 2512
Z.ai · released Dec 9, 2025 · zai-org/GLM-ASR-Nano-2512
Speech to textTranscribes a recording into words
- Type
- Open weightsMIT License
- Languages
- 2
- Size
- 2.3B
Context measured in tokens
Our take
Written Sep 14, 2026GLM ASR Nano 2512 is a 2.3-billion-parameter speech-to-text model from Z.ai with a permissive MIT licence. It transcribes extremely fast and handles podcasts, video and accented audio better than most, though it struggles with clean read-aloud speech and meeting-room recordings.
Pick this for fast batch transcription where speed matters — an hour of audio in roughly eleven seconds on standard hardware. Use it for podcast, video or accented earnings-call audio where it beats most rivals. Skip it if you need clean read-aloud precision, meeting-room transcription, or languages beyond English and Chinese.
The case for it
- Extremely fast: 334× real time on the test hardware used for the leaderboard.
- Strong on podcast and video audio at 8.02% word error rate, better than most of 85 models.
- Handles accented earnings-call audio better than most, at 7.29% word error rate.
- MIT licence and 2.3 billion parameters make it easy to self-host commercially.
The case against it
- Mediocre on clean read-aloud audio at 1.7% word error rate, worse than most.
- Weak on meeting-room audio at 13.95% word error rate, well above the field median.
- Only two languages supported and no commercial hosting available.
How good is it?
An open transcription model for turning speech into text, though meeting recordings are where it struggles most.
- transcribing recordings of meetings in a roomRecorded meetings · 78th of 92
TranscriptionTurning speech into text3 of 5Open ASR WER · 44th of 76
94.7%
Misses roughly one word in 19, averaged over nine English test sets.
334×54th of 74
an hour of audio in 11 seconds, on the board's own hardware. Your machine will differ.
2
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 ↓ ↑
English · Chinese
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 GB spare means a 10% error in the size would not change the answer.
GeForce RTX 5090 · 32 GB
Room to spare. 28 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. 3.2 GB spare means a 10% error in the size would not change the answer.
These cards answer whether GLM ASR Nano 2512 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 allowsMIT License, 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
MIT License
Fully permissive: do anything with attribution. No patent grant, unlike Apache-2.0.
Identifiers
- Hugging Face
- zai-org/GLM-ASR-Nano-2512
- Architecture
- Dense
- Takes in, gives back
- Audio in, text out
- Catalogue slug
- zai-org-glm-asr-nano-2512