Maithra Raghu(@maithra_raghu):New eval results on FrontierFinance: Samaya's System (high effort) reaches 56% accuracy, a new state of the art, at 2x lower cost than Claude Fable 5. Claude Fable 5 scores 49.2%, and Samaya's System (low effort) scores 50.8% at 4x lower cost. More evals coming, including Gemini 3.6 Flash and Claude Opus 5!

Maithra Raghu

@maithra_raghu

Cofounder and CEO @Samaya_AI. Formerly Research Scientist Google Brain (@GoogleAI), PhD in ML @Cornell.

加入 July 2017

542 正在關注 22.5K 粉絲

Maithra Raghu@maithra_raghu

2026.07.28 15:03

New eval results on FrontierFinance: Samaya's System (high effort) reaches 56% accuracy, a new state of the art, at 2x lower cost than Claude Fable 5. Claude Fable 5 scores 49.2%, and Samaya's System (low effort) scores 50.8% at 4x lower cost. More evals coming, including Gemini 3.6 Flash and Claude Opus 5!