Artificial Analysis tests show that Astra and Fable each have their own strengths and weaknesses, with neither establishing an overall lead. Specifically, Astra leads in the GDP.pdf test, while Fable leads in net score for factual Q&A penalizing wrong answers, and neither has secured a victory in real-time research or reliable citation.