Axiom, a startup founded in 2025, has achieved a significant milestone by solving 12 problems on the prestigious Putnam undergraduate math exam, with an AI system scoring 8/12. This accomplishment surpasses other AI systems and approaches the performance of top human undergraduates, highlighting AI's growing capabilities in complex academic challenges. Axiom's approach, termed "Verified AI," focuses on formally proving mathematical concepts using tools like Lean and their own AXLE toolkit, aiming to scale and compound brilliance by building on rigorous foundations. AI
IMPACT Demonstrates AI's advancing capabilities in complex symbolic reasoning and problem-solving, potentially influencing future AI development and evaluation methods.
RANK_REASON AI system achieves a high score on a prestigious academic benchmark. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →