Model comparison

VibeVoice-Large vs Voxtral-Mini-4B-Realtime-2602

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

From the file· summed bytes, KV per layer

Side by side

VibeVoice-LargeVoxtral-Mini-4B-Realtime-2602
Parameters9.3B4.4B
Architecturevibevoice-asrvoxtral_realtime
Layers26
Native context131,072
Mixture of expertsnono
Quantizations published78
Smallest quantization5.74 GiB2.35 GiB
Q4_K_M2.64 GiB
Licencemitapache-2.0

KV cache by context

the term that decides long-context viability
ContextVibeVoice-LargeVoxtral-Mini-4B-Realtime-2602Ratio
4,096
8,192
16,384
32,768
65,536
131,072