Very long inputs
Choose Embed v4.0
Embed v4.0 reads up to 128k tokens per request, against 8k for Codestral Embed.
Choose Codestral Embed for the latest release. Choose Embed v4.0 for very long inputs and images as input.
Each answer below comes straight from the published prices, specifications and independent benchmark results.
Choose Embed v4.0
Embed v4.0 reads up to 128k tokens per request, against 8k for Codestral Embed.
Choose Embed v4.0
Embed v4.0 accepts images as input; Codestral Embed does not.
Choose Codestral Embed
Codestral Embed is the newer release (May 2025, against April 2025 for Embed v4.0).
| Codestral Embed | Embed v4.0 | |
|---|---|---|
| Provider | Mistral AI | Cohere |
| Category | Embeddings | Embeddings |
| Status | Generally available | Legacy |
| Headline price | No data | No data |
| Context window | 8k | 128k |
| Max output | No data | No data |
| Released | May 2025 | April 2025 |
| Knowledge cutoff | No data | No data |
| Licence | Proprietary | Proprietary |
| Capabilities | No data | Vision |
| API model id | codestral-embed-2505 | embed-v4.0 |
Choose Codestral Embed for the latest release. Choose Embed v4.0 for very long inputs and images as input.
A price comparison is not possible: one of them has no comparable published price.
Embed v4.0 reads more per request: 128k tokens against 8k.
Codestral Embed is made by Mistral AI and Embed v4.0 by Cohere. A price comparison is not possible: one of them has no comparable published price. Embed v4.0 reads more per request: 128k tokens against 8k. Codestral Embed is the newer model (May 2025, against April 2025). Only Embed v4.0 offers vision.
Codestral Embed was released on May 2025 and Embed v4.0 on April 2025.