Model comparison

GLM-ASR-Nano-2512 vs whisper-large-v3

These two publish different quantization sets; the table below has the exact sizes.

From the file· summed bytes, KV per layer

Side by side

GLM-ASR-Nano-2512whisper-large-v3
Parameters2.3B1.5B
Architectureglmasrwhisper
Layers2832
Native context8,192
Mixture of expertsnono
Quantizations published518
Smallest quantization0.91 GiB0.49 GiB
Q4_K_M0.93 GiB
Licencemitapache-2.0

KV cache by context

the term that decides long-context viability
ContextGLM-ASR-Nano-2512whisper-large-v3Ratio
4,0960.22 GiB
8,1920.44 GiB
16,3840.88 GiB
32,7681.75 GiB
65,5363.50 GiB
131,0727.00 GiB