Model comparison
Teuken-7B-instruct-research-v0.4 vs gemma-4-E4B-it
At Q4_K_M, gemma-4-E4B-it is the smaller download — 4,977,171,584 bytes against 5,018,868,320. At long context the gap widens: gemma-4-E4B-it's KV cache at 32K is 2.0× smaller, which usually matters more than the difference in weights.
From the file· summed bytes, KV per layer
Side by side
| Teuken-7B-instruct-research-v0.4 | gemma-4-E4B-it | |
|---|---|---|
| Parameters | 7.5B | 8.0B |
| Architecture | llama | gemma4 |
| Layers | 32 | 42 |
| Native context | 4,096 | 131,072 |
| Mixture of experts | no | no |
| Quantizations published | 41 | 36 |
| Smallest quantization | 2.48 GiB | 3.30 GiB |
| Q4_K_M | 4.67 GiB | 4.64 GiB |
| Licence | other | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | Teuken-7B-instruct-research-v0.4 | gemma-4-E4B-it | Ratio |
|---|---|---|---|
| 4,096 | 0.13 GiB | 0.12 GiB | 1.02× |
| 8,192 | 0.25 GiB | 0.18 GiB | 1.41× |
| 16,384 | 0.50 GiB | 0.29 GiB | 1.74× |
| 32,768 | 1.00 GiB | 0.51 GiB | 1.98× |
| 65,536 | 2.00 GiB | 0.94 GiB | 2.12× |
| 131,072 | 4.00 GiB | 1.82 GiB | 2.20× |