| SPECIFICATIONS | |||||
| Max dimension | 4K | 1K | 2K | 2K | |
| Max tokens | 33K | 8K | 128K | 128K | |
| Parameters | 8B | 559M | — | — | |
| License | Apache 2.0 | CC BY-NC 4.0 | Proprietary | Proprietary | |
| EMBEDDING TYPES | |||||
| Types |
Dense
|
Dense
|
Dense
|
Dense
|
|
| MATRYOSHKA DIMENSIONS | |||||
| Available sizes |
3264128256512102420484096
|
32641282565121024
|
25651210241536
|
25651210241536
|
|
| INPUT MODALITIES | |||||
| Text | ✓ | ✓ | ✓ | ✓ | |
| Image | — | — | ✓ | ✓ | |
| — | — | ✓ | ✓ | ||
| OUTPUT TYPES | |||||
| Single vector | ✓ | ✓ | ✓ | ✓ | |
| Multi vector | — | — | — | — | |
| LANGUAGE SUPPORT | |||||
| Languages | 🌍 100+ | 🌍 89+ | 🌍 100+ | 🌍 100+ | |
| PRICING | |||||
| Per 1M tokens | — | $0.020 | — | — | |
| BENCHMARKS | |||||
| CMTEB CHINESE | 73.84 | — | — | — | |
| LONGEMBED | — | 55.66 | — | — | |
| MMTEB | — | 58.58 | — | — | |
| MTEB EN | — | 54.33 | — | — | |
| MTEB EN V2 | 75.22 | — | — | — | |
| MTEB MULTILINGUAL | 70.58 | — | — | — | |