The news: Meta updated its Adaptive Ranking Model for Instagram, allowing its ad serving process to deliver more effective, relevant ads to users with less compute, per a company announcement.

• The revised system processes more user engagement signals in real time while lowering system load to improve the delivery of relevant content.
• Meta said the model now uses LLM-scale intelligence for ad targeting, boosting ad performance and return on ad spend while remaining computationally efficient.

Meta pointed to promising results from its Adaptive Ranking Model after its initial launch in Q4 2025. The company says that the model has so far offered a 5% boost in ad clickthrough rates and a 3% increase in conversions.

Zooming out: Meta is pushing into more efficient, automated ad targeting.

The company introduced options to consolidate ad targeting in 2025, shifting advertising away from its manual roots and toward a future led by Meta’s AI systems—systems it cites as key drivers of increased ad conversions. Its Advantage suite gives advertisers the option to use Meta’s systems as either a fully automated, end-to-end solution, or as modular components that automate where ads appear while giving marketers control over targeting.