GigaChat Embeddings 2 is GigaChat (Sber)'s embedding model with a 512 context window. The second-generation GigaChat embedding model from Sber, producing improved dense text representations for Russian-language retrieval and similarity applications.
gigachat-embeddings-2 |
| Embedding |
| Active |
| 512 tokens |
| Text |
| Embedding |
| 1024 |
Capabilities
Input1/5
✓
·
·
·
·
Output1/5
·
·
·
·
✓
Capabilities0/13
·
·
·
·
·
·
·
·
·
·
·
·
·
Versions
| Version | Released | Context | Input / 1M | Output / 1M | Status |
|---|---|---|---|---|---|
| Text Embedding 5 | 2K | $0.025 | $0.000 | Available | |
| Embed 4 | 128K | $0.120 | $0.470 | Available | |
| Embed 4 Img | — | — | $0.470 | $0.000 | Available |
| Embed 4 Txt | — | — | $0.120 | $0.000 | Available |
| Text Embedding 4 | — | 2K | $0.100 | $0.000 | Deprecated |
| Voyage 4 | — | 32K | $0.060 | $0.000 | Available |
| Voyage 4 Large | — | 32K | $0.120 | $0.000 | Available |
| Voyage 4 Lite | — | 32K | $0.020 | $0.000 | Available |
| Voyage 3.5 | 32K | $0.060 | $0.000 | Available | |
| Voyage 3.5 Lite | 32K | $0.020 | $0.000 | Available | |
| GigaChat Embeddings 2 | — | 512 | — | — | Current |