E5 Mistral 7B Instruct vs Multilingual E5 Large Instruct
E5 Mistral 7B Instruct60
Multilingual E5 Large Instruct63.2
| Benchmark | E5 Mistral 7B Instruct | Multilingual E5 Large Instruct |
|---|---|---|
| MTEB multilingual v2 · Mean over tasks · multilingual (1037 languages) | 60 | 63.2 |
Specs
| Spec | E5 Mistral 7B Instruct | Multilingual E5 Large Instruct |
|---|---|---|
| Price | !$0.022 | !$0.01 |
| Context | 33K | 1K |
| Weights | Open | Open |
| Released | Feb 2024 | Feb 2024 |
| Dimensions | 4,096 | 1,024 |
| Max input | 33K tokens | 1K tokens |
| Languages | en | en, Esperanto, es +90 |
| Memory (4-bit) | 4.3 GB | 0.3 GB |
| Licence | MIT | MIT |
| Commercial use | Allowed | Allowed |
What breaks when you switch
E5 Mistral 7B Instruct → Multilingual E5 Large Instruct
Re-embed your whole corpus›
Vectors from different models can't be mixed; dimensions change from 4,096 to 1,024.
Max input: 33K tokens → 1K tokens
Multilingual E5 Large Instruct → E5 Mistral 7B Instruct
Re-embed your whole corpus›
Vectors from different models can't be mixed; dimensions change from 1,024 to 4,096.