| SPECIFICATIONS | |||||
| Max dimension | 2K | 2K | 1K | 4K | |
| Max tokens | 33K | 33K | 512 | 32K | |
| Parameters | 3.8B | 3.8B | — | 7B | |
| License | Apache 2.0 | Apache 2.0 | Proprietary | Apache 2.0 | |
| EMBEDDING TYPES | |||||
| Types |
DenseLate
|
DenseLate
|
Dense
|
Dense
|
|
| MATRYOSHKA DIMENSIONS | |||||
| Available sizes |
12825651210242048
|
12825651210242048
|
— | — | |
| INPUT MODALITIES | |||||
| Text | ✓ | ✓ | ✓ | ✓ | |
| Image | ✓ | ✓ | ✓ | — | |
| ✓ | ✓ | — | — | ||
| OUTPUT TYPES | |||||
| Single vector | ✓ | ✓ | ✓ | ✓ | |
| Multi vector | ✓ | ✓ | — | — | |
| LANGUAGE SUPPORT | |||||
| Languages | 🌍 29+ | 🌍 29+ | — | 🌍 Multi | |
| PRICING | |||||
| Per 1M tokens | $0.050 | $0.050 | — | — | |
| BENCHMARKS | |||||
| CMTEB 35 | — | — | — | 72.05 | |
| COIR | 71.59 | 71.59 | — | — | |
| JINA VDR | 84.11 | 84.11 | — | — | |
| LONGEMBED | 67.11 | 67.11 | — | — | |
| MMTEB | 66.49 | 66.49 | — | — | |
| MTEB 56 | — | — | — | 70.24 | |
| MTEB AVG | — | — | 64.47 | — | |
| MTEB EN | 55.97 | 55.97 | — | — | |
| MTEB FR | — | — | — | 68.25 | |
| MTEB PL | — | — | — | 67.86 | |
| MTEB RETRIEVAL | — | — | 55.00 | — | |
| VIDORE | 90.17 | 90.17 | — | — | |