Pinterest Achieves 90% AI Cost Reduction Through Blended Open and Closed-Source Models

Pinterest touted a 90% cost reduction in its AI-driven experiences during its February 2026 fourth-quarter earnings presentation, achieved by blending open-source and closed-source generative AI models instead of relying solely on proprietary ones, Business Insider reported.

The social media platform's vice president of product management, Vicky Gkiza, said mixing open-source and closed-source AI models helped cut costs. Pinterest's recent AI-enabled features combine OpenAI, Alibaba, and its own AI models.

To support this approach, Pinterest invested in cloud infrastructure and an updated hiring strategy. Pinterest is taking a model-agnostic approach to generative AI that started in 2023. This approach involves combining proprietary AI models developed by its software engineers, closed-source models from Anthropic and OpenAI, and open-source models from Alibaba, Gkiza said.

Closed-source AI models can quickly process large amounts of data, require little maintenance, and integrate instantly into existing systems due to their pre-built nature. Open-source models are typically free to download, use publicly available data, and can be modified more deeply than closed-source offerings.

However, open-source models require the expertise of software engineers who can build, maintain, and debug these models.

Lan Guan, chief AI and data officer at Accenture, said companies are increasingly seeing value in a multimodal AI approach that balances performance and the cost of tokens. Pinterest bolstered its blended approach to generative AI through beta testing, an updated hiring strategy focused on software development expertise, and infrastructure investments, Gkiza said.

Pinterest uses OpenAI's closed-source large language models to support some of its product features, a company spokesperson said.

Pinterest relies on Anthropic's Claude, a closed-source AI, for internal use cases like coding, the spokesperson added. Alibaba's Qwen, an open-source LLM, is used by Pinterest for visual and content understanding, data labeling, and assistant tasks. Pinterest rolled out several new AI features in 2025, each blending open-source and closed-source generative AI, Gkiza said.

Pinterest began testing auto-collages at the beginning of 2024. The company's software engineers fine-tuned a mix of internally developed, open-source, and third-party AI models for auto-collages in 2024. Auto-collages is a feature where advertisers can convert their product catalogs into pins that populate on shopping feeds for categories like skincare, home decor, and fashion.

By June 2025, auto-collages was ready for a pilot that included a small group of retailer-advertisers like Macy’s. Pinterest’s voice-enabled AI feature underwent beta testing in October 2025. The voice-enabled AI feature uses both open-source and third-party AI to generate responses to users' queries.

Early results from voice-enabled AI beta testing showed that users tended to pose more shopping-specific questions when asking aloud rather than typing searches, a Pinterest spokesperson said. Gkiza said, 'Search has been evolving so fast. Pinterest hired employees with AI and machine learning expertise to steward the customization of its large language models, Gkiza said.

Former Google engineer Matthias Zenger joined Pinterest as its vice president of engineering in April 2025. In August 2025, Pinterest announced it had hired software engineer Mirjam Wattenhofer to focus on e-commerce and user experiences. In August 2025, Pinterest announced it would open an Engineering Excellence Hub in Zurich.

Matthias Zenger and Mirjam Wattenhofer work at Pinterest's Zurich office. The Zurich office team focuses on improving user experiences with AI and machine-learning technologies. In February 2025, Pinterest CEO Bill Ready said the company would hire additional research and development workers to support its AI efforts.

Pinterest's blended approach to AI-driven experiences costs an estimated 90% less than when it relied solely on its proprietary models, Gkiza said. Pinterest plans to invest more in cloud-computing infrastructure like graphics processing units to power its cost-saving AI technology, Gkiza said.

Pinterest will continue to experiment with various AI models, prioritizing its own models for personalization, open-source models for cost-effectiveness and multimodal machine learning, and closed-source models when they offer the best performance.