Granite Speech 3.3 2b
IBM · released Apr 28, 2025 · ibm-granite/granite-speech-3.3-2b
Speech to textTranscribes a recording into words
- Type
- Open weightsApache License 2.0
- Languages
- 5
- Size
- 3B
Context measured in tokens
Our take
Written Sep 4, 2026Granite Speech 3.3 2b is a compact downloadable speech-to-text model from IBM with a permissive Apache licence. It transcribes extremely fast and handles clean, financial and meeting audio well, though podcasts and video are weak spots.
Pick this for fast batch transcription — an hour of audio in about seven seconds. Use it for clean read-aloud, financial calls, meetings or European-accented English, where it beats most rivals. Good for research or commercial self-hosting under Apache 2.0. Skip it if you need hosted inference, are transcribing podcasts or video, or need verified accuracy beyond English.
The case for it
- Transcribes at 507 times real time — an hour of audio in roughly seven seconds on the test harness.
- Strong on clean and semi-clean audio: about one word in eighty wrong on read-aloud, and better than most on financial calls and European-accented speech.
- Resilient on hard audio where most models struggle: better than most on meetings and heavily accented speech.
- Apache 2.0 licence allows commercial use, modification and redistribution without restriction.
The case against it
- Podcasts and video are a clear weak spot at 9.0% word error, worse than most in the field.
- No hosted providers currently listed; you must run it yourself.
- Five languages are listed, but every accuracy figure we hold is English only.
How good is it?
An open speech-to-text model for turning recordings of meetings into written text.
- transcribing recordings of meetings in a roomRecorded meetings · 18th of 92
TranscriptionTurning speech into text3 of 5Open ASR WER · 45th of 76
94.6%
Misses roughly one word in 19, averaged over nine English test sets.
509×47th of 74
an hour of audio in 7 seconds, on the board's own hardware. Your machine will differ.
5
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 · French · German · Spanish · Portuguese
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. 19.5 GB spare means a 10% error in the size would not change the answer.
GeForce RTX 5090 · 32 GB
Room to spare. 27.5 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. 2.7 GB spare means a 10% error in the size would not change the answer.
These cards answer whether Granite Speech 3.3 2b 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.
Models people weigh against Granite Speech 3.3 2b
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
- ibm-granite/granite-speech-3.3-2b
- Architecture
- Dense
- Takes in, gives back
- Audio in, text out
- Catalogue slug
- ibm-granite-granite-speech-3-3-2b