Samaya AI's Sterling Beats Claude on Finance Research at Lower Cost
Samaya AI has released Sterling, a model built for financial research. It scored 58.9% on a finance test called FrontierFinance, ahead of Claude Fable 5.1 at 55.9%. It also costs much less per question: $0.61 versus $6.89. The model uses about 3 billion active parameters, a measure of its size.
Quick glossarySamaya AI, FrontierFinance, 3B-Active, Nemotron
- Samaya AI
- A company that built the Sterling model; the story gives no further background about it.
- FrontierFinance
- A test that measures how accurately AI systems answer tough questions in the finance research field.
- 3B-Active
- The parts of an AI model's internal settings that actually work on each task, a rough sign of its size.
- Nemotron
- A family of AI models from Nvidia, the chipmaker, used here as the base for Sterling.
Burning Potato score: 73%