Bekko Embedding v1 A25m vs Bekko Embedding v1 A8m
MTEB multilingual v2 · Mean over tasks · multilingual (1037 languages) Embeddings assign examples to categories, group similar examples, estimate semantic similarity between pairs, classify relationships between pairs, rank relevant documents for queries, find translated sentence pairs, rerank documents according to instructions, assign multiple labels to each example, reorder candidate documents by relevance on MTEB Multilingual v2 · mean over tasks in 1037 languages including 'Auhelawa, Abau, Achi, Achinese, Achuar-Shiwiar, Adzera, Afrikaans, Agarabi; tasks receive equal weight
Bekko Embedding v1 A25m58.4
Bekko Embedding v1 A8m56.7
| Benchmark | Bekko Embedding v1 A25m | Bekko Embedding v1 A8m |
|---|---|---|
| MTEB multilingual v2 · Mean over tasks · multilingual (1037 languages) | 58.4 | 56.7 |
Specs
| Spec | Bekko Embedding v1 A25m | Bekko Embedding v1 A8m |
|---|---|---|
| Context | 8K | 8K |
| Weights | Open | Open |
| Released | Jul 2026 | Jul 2026 |
| Dimensions | 384 | 384 |
| Max input | 8K tokens | 8K tokens |
| Memory (4-bit) | 0.1 GB | 0.1 GB |
| Licence | MIT | MIT |
| Commercial use | Allowed | Allowed |
What breaks when you switch
Bekko Embedding v1 A25m → Bekko Embedding v1 A8m
Re-embed your whole corpus›
Vectors from different models can't be mixed.
Bekko Embedding v1 A8m → Bekko Embedding v1 A25m
Re-embed your whole corpus›
Vectors from different models can't be mixed.