The article introduces Thomson, a general-purpose frontier model developed by SovereignAI that demonstrates how institutions can achieve frontier performance through continual learning on readily available open-weight models. This approach utilizes a modern mid- and post-training stack with safeguards to preserve plasticity and stability, allowing for minimal high-impact interventions on parameters.
- Thomson performs competitively with recent frontier models across agentic tasks, safety, legal, tax, multilingualism, and large-scale Deep Research.
- Evaluations reveal a distinctive π-shaped pattern of improvements across a wide range of capabilities, including those not explicitly targeted.
- The method almost completely eliminates the forgetting problem common to narrow domain adaptation.
- The approach requires compute and personnel budgets substantially lower than commonly thought, making ownership of the SovereignAI stack viable for more actors.
By exploiting continual learning rather than limited approaches like fine-tuning or prompt engineering, this work argues that frontier performance is achievable for a wide range of institutions under diverse funding settings.