● Covered by 2 sources · 2 reportsMedium impact1 neutral1 positive
Thomson Reuters Develops Proprietary AI Model for Legal and Tax Work
Key points
- Thomson Reuters developed an AI model for legal, tax, and compliance.
- The model was trained on proprietary content for approximately $40 million.
- It will power features like Tabular Analysis in CoCounsel Legal.
- Thomson Reuters still uses third-party models for other product features.
- Thomson Reuters' proprietary LLM is named "Thomson".
- The model was launched on August 24, 2026.
- Thomson Reuters used an open-source foundation to train Thomson.
Proprietary AI Development
Thomson Reuters has developed its own AI model, named Thomson, specifically for legal, tax, and compliance work. This model was trained using the company's extensive proprietary content, including data from Westlaw, Practical Law, Checkpoint, and Reuters. The development involved an investment of approximately $40 million, covering compute resources and talent.
Training Approach and Cost
Instead of building a foundation model from scratch, Thomson Reuters started with an existing open-source foundation. The majority of the investment focused on further training the model with decades of proprietary content and expert-driven evaluation. This strategy offers an alternative for companies with large datasets, allowing them to create specialized AI without the multi-billion dollar cost of developing a new foundation model.
Integration and Use Cases
The Thomson AI model is designed to power specific features within Thomson Reuters products. For example, it will be the default model for Tabular Analysis in CoCounsel Legal, a feature capable of processing up to 10,000 documents and answering 100 questions. This capability is aimed at legal, tax, and compliance professionals who use these products daily.
Continued Use of Third-Party Models
Despite developing its own model, Thomson Reuters continues to utilize frontier models from other providers. Its new CoCounsel Legal product, for instance, is built on Anthropic's Claude Agent SDK. This indicates a hybrid approach where specialized in-house models handle specific tasks, while external models are integrated for broader functionalities.
Updates
- Thomson Reuters' proprietary LLM is named "Thomson".
- The model was launched on August 24, 2026.
- Thomson Reuters used an open-source foundation to train Thomson.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
How outlets covered it
Thomson Reuters has launched "Thomson," its first proprietary large language model, developed in-house for $40 million. This model was trained on decades of Thomson Reuters' content and expertise, aiming to provide professional AI at a lower cost than typical frontier models.
Thomson Reuters has developed its own AI model, named Thomson, for legal, tax, and compliance tasks, trained on its proprietary content. This model will power specific features within products like CoCounsel, while the company continues to use third-party models like Anthropic's Claude for other functionalities. This approach allows companies with extensive proprietary data to create specialized AI without building a foundation model from scratch.