Embed v4.0
Generally availableProprietary128K context
Summary
Embed v4.0 is a proprietary language model from Cohere, released on 15 Apr 2025. It accepts text and generates text, with a 128K-token context window and up to 2K output tokens per response.
Multimodal embedding model for text, images, and mixed text-and-image documents
Last verified 21 Oct 2026. Every figure on this page links to its source.
Specifications
| Developer | Cohere |
|---|---|
| Release date | 15 Apr 2025 [models.dev] |
| Status | Generally available [models.dev] |
| Weights | Proprietary (API only) [models.dev] |
| Context window | 128K (128,000 tokens) [models.dev] |
| Max output | 2K tokens [models.dev] |
| Input | Text [models.dev] |
| Output | Text [models.dev] |
| Reasoning mode | No [models.dev] |
| Tool calling | No [models.dev] |
Embed v4.0 API pricing by provider
USD per million tokens. Sorted by input price.
| Provider | Input | Output | Cached input | Context | Source |
|---|---|---|---|---|---|
| Vercel AI Gateway cohere/embed-v4.0 | – | – | – | 128K | models.dev, 21 Oct 2026 |
Frequently asked questions
What is the context window of Embed v4.0?
Embed v4.0 has a context window of 128,000 tokens (128K) and can generate up to 1,536 tokens in a single response, according to models.dev.
Is Embed v4.0 open source?
No. Embed v4.0 is a proprietary model; Cohere has not released its weights. It is available through APIs including Vercel AI Gateway.
When was Embed v4.0 released?
Cohere released Embed v4.0 on 15 Apr 2025, according to models.dev.
Change history
- Added to the directory
- Now listed on Vercel AI Gateway