Cohere: Embed v4.0
Embed v4.0 is an embedding model from Cohere released on Apr 15, 2025. It reads text and writes vectors, with a 128K token context window.
- Modalities
- In / out price
- n/a
- Context
- 128K
- Max output
- 2K
- Released
- Apr 15, 2025
- Knowledge cutoff
- n/a
- Tool calling
- Structured output
- Reasoning
- Prompt caching
- Open weights
- Free to use
Supported request parameters
What you can set when you call Embed v4.0 through an OpenAI compatible API.
- Temperature
temperature
More from Cohere
Embed 5 Fast
Embeddings128K contextTurns text into vectors for search and retrieval.
by cohereSep 30, 2026
Embed 5 Pro
Embeddings128K contextTurns text into vectors for search and retrieval.
by cohereSep 30, 2026
- 33K context
Reads text.
by cohereSep 9, 2026$0/M input tokens$0/M output tokens
- 256K context
Reads text. Tool calling, reasoning.
by cohereJun 17, 2026$0/M input tokens$0/M output tokens
- Index 9.9
Reads text. Tool calling, structured output, reasoning.
by cohereJun 9, 2026256K context$0/M input tokens$0/M output tokens
- Index 13.1
Reads text and images. Tool calling, structured output, reasoning. Served by 2 providers.
by cohereMay 20, 2026192K context$0.30/M input tokens$1.50/M output tokens2 providers
Embed v4.0 questions
What is the context window of Embed v4.0?
128,000 tokens of input, and it can write up to 1,536 tokens in one reply. As a rough guide, a million tokens is around 750,000 English words.
Is Embed v4.0 open source?
No. Cohere serves it through an API only, so you cannot download the weights or host it yourself.
When was Embed v4.0 released?
Apr 15, 2025.
Catalog data from models.dev (MIT licensed), refreshed daily, last on Oct 15, 2026. The summary is written from these facts.
Building on Embed v4.0?
I test your real tasks on it and the two closest alternatives, then ship the agent on whichever wins on cost and accuracy.
Book Meeting