Codestral Embed is designed for use cases such as code completion, editing, or explanation tasks. It can also be applied in semantic search, duplicate detection, and repository-level analytics across large-scale codebases, the company said.
“Codestral Embed supports unsupervised grouping of code based on functionality or structure,” Mistral AI added. “This is useful for analyzing repository composition, identifying emergent architecture patterns, or feeding into automated documentation and categorization systems.”
The model is available through Mistral’s API under the name codestral-embed-2505, priced at $0.15 per million tokens. A batch API version is offered at a 50 percent discount, and on-premise deployments are available through direct consultation with the company’s applied AI team.
The launch follows Mistral’s recent introduction of the Agents API, which the company said complements its Chat Completion API and is intended to simplify the development of agent-based applications.
Enterprise interest in embeddings
Advanced code embedding models are gaining traction as key tools in enterprise software development, offering improvements in productivity, code quality, and risk management across the software lifecycle.