| SPECIFICATIONS | |||||
| Max dimension | — | 4K | 1K | 2K | |
| Max tokens | — | 33K | 512 | 32K | |
| Parameters | 350M | 8B | — | — | |
| License | Apache 2.0 | Apache 2.0 | Proprietary | Proprietary | |
| EMBEDDING TYPES | |||||
| Types |
LateColBERT
|
Dense
|
Dense
|
Dense
|
|
| MATRYOSHKA DIMENSIONS | |||||
| Available sizes | — |
3264128256512102420484096
|
— |
25651210242048
|
|
| INPUT MODALITIES | |||||
| Text | ✓ | ✓ | ✓ | ✓ | |
| Image | — | — | ✓ | — | |
| — | — | — | — | ||
| OUTPUT TYPES | |||||
| Single vector | — | ✓ | ✓ | ✓ | |
| Multi vector | ✓ | — | — | — | |
| LANGUAGE SUPPORT | |||||
| Languages | 🌍 Multi | 🌍 100+ | — | 🌍 Multi | |
| PRICING | |||||
| Per 1M tokens | — | — | — | — | |
| BENCHMARKS | |||||
| CMTEB CHINESE | — | 73.84 | — | — | |
| MTEB AVG | — | — | 64.47 | — | |
| MTEB EN V2 | — | 75.22 | — | — | |
| MTEB MULTILINGUAL | — | 70.58 | — | — | |
| MTEB RETRIEVAL | — | — | 55.00 | — | |