Harmonic's Aristotle model achieved gold medal performance at the International Mathematical Olympiad (IMO)—solving 5 of 6 problems—using a fundamentally different training approach from large language models: converting math into verifiable code, machine-checking proofs automatically, and using synthetic self-generated data to train via reinforcement learning rather than internet-scraped text.
factualpending
Speaker
Vlad TenevEvidence Quote
“earlier this year, we actually achieved gold medal performance at the International Mathematical Olympiad...Aristotle, which is our model, solved five out of six.”
Created: 8/10/2026, 11:02:14 PM
Your notes
Loading notes...