Very long inputs
Choose Timestral 1.0
Timestral 1.0 reads up to 32k tokens per request, against 16k for North Small Translate.
Choose Timestral 1.0 for very long inputs and self-hosting and data control; on the other points the two are level or the data is missing.
Each answer below comes straight from the published prices, specifications and independent benchmark results.
Choose Timestral 1.0
Timestral 1.0 reads up to 32k tokens per request, against 16k for North Small Translate.
Choose Timestral 1.0
Timestral 1.0 has open weights, so you can run it on your own servers; North Small Translate is only available through its provider.
| North Small Translate | Timestral 1.0 | |
|---|---|---|
| Provider | Cohere | Mistral AI |
| Category | Specialised | Specialised |
| Status | Generally available | Preview |
| Headline price | No data | No data |
| Context window | 16k | 32k |
| Max output | No data | No data |
| Released | September 2026 | No data |
| Knowledge cutoff | No data | No data |
| Licence | Proprietary | Open weights |
| Capabilities | No data | Self-hostable |
| API model id | north-small-translate-09-2026 |
Choose Timestral 1.0 for very long inputs and self-hosting and data control; on the other points the two are level or the data is missing.
A price comparison is not possible: one of them has no comparable published price.
Timestral 1.0 reads more per request: 32k tokens against 16k.
North Small Translate is made by Cohere and Timestral 1.0 by Mistral AI. A price comparison is not possible: one of them has no comparable published price. Timestral 1.0 reads more per request: 32k tokens against 16k. Only Timestral 1.0 offers self-hostable.