nexos.ai Launches Smart Router to Slash AI Coding Costs by 60%
News related to:nexos.ai · 2 min read
nexos.ai, a Lithuanian AI infrastructure company, has launched a smart router that slashes AI coding costs by 60% for engineering teams. The new product addresses a growing challenge in the industry: while the cost of individual AI tokens has dropped, the overall volume of AI traffic has surged, making AI spending the fastest-growing expense in engineering budgets.
The smart router automatically matches engineering tasks to the most cost-effective AI models. For complex tasks, it routes them to frontier models, while routine work is handled by low-cost alternatives. This approach ensures that the most expensive models are only used where they add the most value, thereby reducing overall costs without disrupting developer workflows.
According to Žilvinas Girėnas, head of product at nexos.ai, the challenge lies in optimizing AI spending.
To address this issue, nexos.ai developed a novel benchmarking method called Mirror benchmarking. This method continuously evaluates live production traffic at the session level, providing real-time insights into cost-to-quality ratios. The platform uses these insights to make intelligent routing decisions, ensuring that tasks are matched to the most appropriate models.
In production testing, the smart router demonstrated significant cost savings. On one workload, it cut costs by 59.2%, saving over $5,400 on traffic that would have cost more than $9,200 at frontier-model prices. Another test achieved a 60.4% reduction in costs. The quality remained intact because the frontier model still made every critical decision.
The smart router preserves the tiered structure that coding agents already use internally. It routes 16% of requests to frontier reasoning models for complex tasks, while 84% of requests are routed to cost-efficient, open-weight models for routine work. This approach ensures that the most expensive models are only used where they add the most value, thereby reducing overall costs.
By utilizing the smart router, enterprises can reduce their dependency on any single AI provider. The router reads each request without altering it and switches models only at natural breakpoints in the session, keeping the cache largely intact to prevent costly rebuilds. This ensures that the router holds up against the unpredictable reality of production.
Founded in 2024 by Tomas Okmanas and Eimantas Sabaliauskas, who also co-founded the $3B cybersecurity unicorn Nord Security and Oxylabs, nexos.ai addresses the urgent need for enterprises to efficiently deploy, manage, and optimize AI models within organizations. The company's platform provides a centralized solution for seamlessly integrating and managing multiple AI models, ensuring that enterprises can make the most of their AI investments while keeping costs under control.