Cohere and RWS launch Language Weaver Pro for enterprise AI translation

What launched

Cohere said on March 25, 2026 that its frontier models now power RWS Language Weaver Pro, a new enterprise translation offering aimed at organizations that need secure and accurate multilingual workflows. The announcement on X framed the release around high-stakes use cases for enterprises and governments, while RWS's launch announcement described Language Weaver Pro as a new enterprise-grade AI translation solution built in partnership with Cohere.

RWS says the product ranked first in 31 of 32 languages in its benchmarking tests and outperformed DeepL and Gemini across sentence-level and paragraph-level evaluations. The company also says Language Weaver Pro is a 100+ billion parameter model, calling it the largest dedicated translation model currently in production. RWS positions the system as designed for business-critical content where accuracy, governance, and domain sensitivity matter more than raw consumer convenience.

How RWS positions the product

The RWS release argues that many AI translation systems can produce fluent language but still miss meaning, culture, or compliance requirements. Language Weaver Pro is intended to close that gap with better contextual understanding while keeping enterprise security and governance requirements intact. The broader Language Weaver product site emphasizes secure, scalable AI translation for sensitive operational environments and points to regulated sectors such as government, legal, healthcare, and other enterprise workflows.

RWS also says Language Weaver Pro is integrated natively across the Trados portfolio, which matters for buyers already running established localization pipelines. That means the launch is not just a model announcement. It is also a packaging and workflow story about how a stronger translation engine gets inserted into existing enterprise tooling.

Why it matters

This is a notable move for Cohere because it extends the company's enterprise AI footprint into a translation stack with explicit benchmarking claims against major rivals. It is also notable for RWS because the company is tying its next translation tier to frontier-model branding, large-model scale, and regulated-environment positioning all at once. For global organizations, the more important question will be whether those benchmark wins translate into lower human review cost without sacrificing security or style control.