Closed-weight speech models
16
Models
9 organisations
0%
Open weights
0 open, 16 closed
0
Frontier
By Epoch's flag
1B
Largest traceable
Conformer + Wav2vec 2.0 + Noisy Student
2009–2023
Span
7 with a traceable size
Organisation
Records
16| Model | Organisation | Released | Parameters | Weights |
|---|---|---|---|---|
| Mamba-24M (SC09) | Carnegie Mellon University (CMU) | 2023-12 | 23M | Closed |
| SauTech | Saudi Data and Artificial Intelligence Authority | 2022-09 | — | Closed |
| W2v-BERT | Google Brain | 2021-08 | 1B | Closed |
| top-down frozen classifier | University of Edinburgh | 2021-02 | — | Closed |
| BigSSL | 2021-01 | 8B * | Closed | |
| Conformer + Wav2vec 2.0 + Noisy Student | 2020-10 | 1B | Closed | |
| Conformer | 2020-05 | 119M | Closed | |
| ContextNet | 2020-05 | 113M * | Closed | |
| ContextNet + Noisy Student | 2020-01 | — | Closed | |
| DL scaling speech | Baidu | 2017-12 | 193M * | Closed |
| LF-MMI | Johns Hopkins University | 2016-09 | 17M | Closed |
| MS-ensemble-speech-recognition | Microsoft | 2016-09 | 3B * | Closed |
| Segmental RNN | University of Edinburgh | 2016-06 | — | Closed |
| TC-DNN-BLSTM-DNN | Carnegie Mellon University (CMU) | 2015-04 | 18M | Closed |
| Dropout (TIMIT) | University of Toronto | 2012-06 | 49M * | Closed |
| BP-DBN | University of Toronto | 2009-01 | 18M | Closed |
An asterisk marks a parameter figure the source does not consider traceable to the people who built the model. That judgement is the source’s and it has false negatives: Kimi K3 is marked speculative at 2.8T while Moonshot’s own model card states the figure. Why that distinction carries the whole exhibit.
Counts are of models Epoch AI records as notable, not of every model released. The most recent month is always incomplete, because a model enters the dataset when it is reviewed rather than when it ships.
Epoch AI, 'Data on AI Models'. Published online at epoch.ai. Retrieved from 'https://epoch.ai/data/ai-models-documentation' Used under CC BY 4.0.