Model comparison
whisper-large vs cohere-transcribe-03-2026
At Q4_K_M, whisper-large is the smaller download — 996,526,048 bytes against 1,558,162,944.
From the file· summed bytes, KV per layer
Side by side
| whisper-large | cohere-transcribe-03-2026 | |
|---|---|---|
| Parameters | 1.5B | 2.1B |
| Architecture | whisper | cohere_asr |
| Layers | 32 | — |
| Native context | — | — |
| Mixture of experts | no | no |
| Quantizations published | 6 | 11 |
| Smallest quantization | 0.93 GiB | 1.41 GiB |
| Q4_K_M | 0.93 GiB | 1.45 GiB |
| Licence | apache-2.0 | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | whisper-large | cohere-transcribe-03-2026 | Ratio |
|---|---|---|---|
| 4,096 | — | — | — |
| 8,192 | — | — | — |
| 16,384 | — | — | — |
| 32,768 | — | — | — |
| 65,536 | — | — | — |
| 131,072 | — | — | — |