Thomson Reuters Launches Its $40M Internal AI Based on Qwen
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
Thomson Reuters has chosen to develop its own language model, named Thomson and based on Alibaba's Qwen, at a cost of approximately $40 million over two years. The best results are achieved when the model has access to proprietary content, particularly Westlaw.
Performance Depends on Access to Internal Content
Performance tests show that the model achieves its best results when it has access to internal content, such as Westlaw. Joel Hron, Chief Technology Officer, emphasizes that the key is not the intelligence itself, but determining which intelligence is relevant to possess.
An Internal Model Based on Qwen
Thomson Reuters has developed an internal language model called Thomson, built from Qwen, Alibaba's model.
A $40 Million Investment Over Two Years for Proprietary AI
The project's cost amounts to approximately $40 million over two years. This choice allows Thomson Reuters to own its AI, rather than renting models offered by OpenAI or Anthropic.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.