Model comparison
Olmo-3.1-32B-Instruct vs gemma-4-E4B-it
At Q4_K_M, gemma-4-E4B-it is the smaller download — 4,977,171,584 bytes against 19,482,034,976. At long context the gap widens: gemma-4-E4B-it's KV cache at 32K is 5.6× smaller, which usually matters more than the difference in weights.
From the file· summed bytes, KV per layer
Side by side
| Olmo-3.1-32B-Instruct | gemma-4-E4B-it | |
|---|---|---|
| Parameters | 32.2B | 8.0B |
| Architecture | olmo2 | gemma4 |
| Layers | 64 | 42 |
| Native context | 65,536 | 131,072 |
| Mixture of experts | no | no |
| Quantizations published | 49 | 36 |
| Smallest quantization | 6.75 GiB | 3.30 GiB |
| Q4_K_M | 18.14 GiB | 4.64 GiB |
| Licence | apache-2.0 | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | Olmo-3.1-32B-Instruct | gemma-4-E4B-it | Ratio |
|---|---|---|---|
| 4,096 | 1.00 GiB | 0.12 GiB | 8.13× |
| 8,192 | 1.34 GiB | 0.18 GiB | 7.56× |
| 16,384 | 1.84 GiB | 0.29 GiB | 6.42× |
| 32,768 | 2.84 GiB | 0.51 GiB | 5.62× |
| 65,536 | 4.84 GiB | 0.94 GiB | 5.13× |
| 131,072 | 8.84 GiB | 1.82 GiB | 4.86× |