Custom Models for Law, News, and Finance: Thomson Reuters’ Thomson LLM, based on Qwen 3.5, trained on proprietary data

modelsAndrew NgSep 4, 2026

Thomson Reuters launched a proprietary large language model family called Thomson.

Impact: LowConfidence: LowNovelty: Incremental
modelssource-feed-e3a85c19-8c35-4364-ad65-15026e2446cb

See the whole story

Read the original on Andrew Ng

Sens.ai aggregates and assesses published reporting. The assessment above is machine generated; the original source is authoritative.