Model comparison
Muse-Glimmer-30B vs gemma-4-26B-A4B-it
At Q4_K_M, gemma-4-26B-A4B-it is the smaller download — 16,796,017,248 bytes against 17,306,324,000. At long context the gap widens: Muse-Glimmer-30B's KV cache at 32K is 3.1× smaller, which usually matters more than the difference in weights.
From the file· summed bytes, KV per layer
Side by side
| Muse-Glimmer-30B | gemma-4-26B-A4B-it | |
|---|---|---|
| Parameters | 29.8B | 26.5B |
| Architecture | muse-glimmer | gemma4 |
| Layers | 52 | 30 |
| Native context | 131,072 | 262,144 |
| Mixture of experts | no | yes, 128 experts |
| Quantizations published | 36 | 46 |
| Smallest quantization | 8.31 GiB | 8.99 GiB |
| Q4_K_M | 16.12 GiB | 15.64 GiB |
| Licence | apache-2.0 | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | Muse-Glimmer-30B | gemma-4-26B-A4B-it | Ratio |
|---|---|---|---|
| 4,096 | 0.15 GiB | 0.45 GiB | 3.08× |
| 8,192 | 0.20 GiB | 0.61 GiB | 3.08× |
| 16,384 | 0.30 GiB | 0.92 GiB | 3.08× |
| 32,768 | 0.50 GiB | 1.54 GiB | 3.08× |
| 65,536 | 0.91 GiB | 2.79 GiB | 3.08× |
| 131,072 | 1.72 GiB | 5.29 GiB | 3.08× |