NEAR AI Proves 672 Putnam Problems for Just $111
NEAR AI’s open-source Lean agent has solved all 672 problems in the PutnamBench theorem-proving benchmark for a reported total cost of $111, according to NEAR Protocol co-founder Alex Skidanov. The result is approximately 250 times cheaper than the second-lowest known submission and significantly below the estimated $10,000–$25,000 cost of competing full-benchmark runs.
PutnamBench uses Lean 4 to require machine-verifiable mathematical proofs, making it a demanding test for AI reasoning and formal verification. NEAR AI’s achievement highlights major improvements in agent efficiency and could lower the cost of developing and testing AI theorem-proving systems.
The NEAR AI agent is open source, allowing researchers and developers to inspect, adapt and extend the technology. The project links the result to NEAR’s broader IronClaw vision for confidential and verifiable AI. For crypto traders, the announcement is primarily a technology and ecosystem signal rather than a direct market catalyst. It may support long-term interest in NEAR’s AI infrastructure narrative, but the article provides no new token utility, funding, network-usage or protocol-growth data.
Neutral
The expected market impact is neutral because the announcement demonstrates a technical achievement but does not directly change NEAR’s token economics, network activity, revenue, staking demand or protocol functionality. The 250-fold cost advantage and open-source release could strengthen NEAR’s AI infrastructure narrative and attract developers over the long term. However, traders typically require measurable adoption, partnerships, funding or token-related catalysts before repricing an asset substantially.
In the short term, NEAR could receive limited speculative attention, particularly if the broader market is favouring AI and crypto infrastructure themes. Any rally would likely depend on trading volume, social-media momentum and confirmation from official ecosystem metrics. Similar announcements involving AI benchmarks, zero-knowledge technology or developer tools have often produced brief narrative-driven moves rather than sustained trends when they lack immediate token utility.
Over the longer term, the project could become more significant if the Lean agent leads to real deployments, developer adoption or integration with NEAR’s confidential-AI initiatives. Conversely, the benchmark result alone does not prove commercial demand or superiority across broader workloads. Traders should therefore treat the news as a sentiment-positive ecosystem development, while monitoring NEAR liquidity, relative strength, on-chain activity and follow-up announcements.