Thomson Reuters Trains Own AI Model
The company continues relying on Anthropic despite training its own model.
TLDR
The New Stack reported that Thomson Reuters trained its own AI model on proprietary legal content at a cost of $40M. The article states the company still kept using Anthropic's model anyway. The source summary notes the model targets legal and tax platforms and competes on domain benchmarks with OpenAI and Anthropic. No first-party confirmation from Thomson Reuters appears in the packet.
Combined views
767
1 Source, first seen ago