A large-scale deployment of 1,000 Fable 5.2 agents successfully formalized Fermat's Last Theorem within 30 hours, translating the complex mathematical proof into a computer-verifiable format. In contrast, the Opus 5.5 model achieved only 26% completion of the same task in 18 hours and showed minimal further progress. This demonstrates a significant performance difference between the two AI models on a rigorous formalization task. AI
IMPACT Highlights significant advancements in AI's capability for complex mathematical reasoning and formal verification.
RANK_REASON Research milestone involving formalization of a mathematical theorem by AI agents. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →