FrontierThe story, in brief
Google’s Aletheia Advances the State of the Art of Fully Autonomous Agentic Math Research
91.9% on IMO-ProofBench. Google's Aletheia just crossed the threshold where AI agents can autonomously discover research-level proofs.
The KeyNews take
Why it matters
Google's Aletheia demonstrates a critical capability milestone: fully autonomous agents solving novel mathematical problems without human intervention. This signals a shift from assisted research tools to independent discovery systems—reshaping how R&D teams approach mathematical problem-solving.
The key facts
5 to knowAletheia solved 6/10 novel math problems in FirstProof challenge
Scored ~91.9% on IMO-ProofBench
Built on Gemini 3 Deep Think
Demonstrates fully autonomous agentic math research without human intervention
FirstProof and IMO-ProofBench are research-level capability benchmarks
Go to the source
InfoQ AI/MLinfoq.com
Publisher excerpt: Google announced Aletheia, an AI using Gemini 3 Deep Think that solved 6/10 novel math problems in the FirstProof challenge. Aletheia also scored ~91.9% on IMO-ProofBench, signaling a significant shift in automated research-level proof discovery without human intervention. By Bruno Couriol