Model comparison
nomic-embed-text-v1.5 vs gte-small
At Q4_K_M, gte-small is the smaller download — 29,203,744 bytes against 84,106,624. At long context the gap widens: gte-small'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
| nomic-embed-text-v1.5 | gte-small | |
|---|---|---|
| Parameters | 137M | 33M |
| Architecture | nomic-bert | bert |
| Layers | 12 | 12 |
| Native context | 2,048 | 512 |
| Mixture of experts | no | no |
| Quantizations published | 39 | 12 |
| Smallest quantization | 0.05 GiB | 0.02 GiB |
| Q4_K_M | 0.08 GiB | 0.03 GiB |
| Licence | apache-2.0 | mit |
KV cache by context
the term that decides long-context viability
| Context | nomic-embed-text-v1.5 | gte-small | Ratio |
|---|---|---|---|
| 4,096 | 0.14 GiB | 0.07 GiB | 2.00× |
| 8,192 | 0.28 GiB | 0.14 GiB | 2.00× |
| 16,384 | 0.56 GiB | 0.28 GiB | 2.00× |
| 32,768 | 1.13 GiB | 0.56 GiB | 2.00× |
| 65,536 | 2.25 GiB | 1.13 GiB | 2.00× |
| 131,072 | 4.50 GiB | 2.25 GiB | 2.00× |