"Large Model Aggregation Platform" OpenRouter is in talks for a new round of funding, with a valuation approaching $1.3 billion.

As AI applications accelerate and evolve toward multi-model architectures, the infrastructure platform helping developers efficiently orchestrate and select large language models is drawing growing investor attention.

On April 1, according to technology media The Information, two people familiar with the matter said that AI model aggregation platform OpenRouter is in talks for a new round of financing totaling $120 million, led by a venture capital arm within Alphabet, Google’s parent company. After the financing is completed, the company’s valuation (including this round of investment) will be about $1.3 billion.

This fundraising development signals that since the company was founded in 2023, it has made the leap from startup to unicorn level in less than two years.

On the revenue front, OpenRouter’s current annualized revenue has already exceeded $50 million. That represents roughly a fivefold increase compared with the more than $10 million annualized revenue disclosed in October last year—its growth rate is significant. This figure reflects the rapid expansion of developers’ demand for multi-model orchestration services, and provides a solid fundamental support for this funding round.

Core product: an API to connect 300+ models

The report says OpenRouter’s core value proposition lies in simplifying the path for developers to access AI models.

With a single API interface, developers can access more than 300 open-source and closed-source models from mainstream vendors such as OpenAI, Anthropic, and Google, and compare the quality, pricing, and response speed of different models on the same platform—so they can decide which model best powers their applications, or assign different models to handle different tasks within the application.

This model aligns with a key trend in AI application development: more and more AI applications and agents no longer rely on a single model. Instead, they dynamically call multiple models based on the task type.

**Microsoft’s actions this week confirm the trend: **the company is rolling out new features for its Office 365 Copilot assistant, enabling users to call OpenAI models to compile materials, call Anthropic models to verify and recheck, and add side-by-side comparison functionality to compare different AI providers’ response results.

Business model: usage-based pricing; annualized revenue grows fivefold

OpenRouter’s path to profitability is relatively straightforward—when developers purchase credits to run AI models, the platform charges a small transaction fee.

According to the report, people familiar with the matter said the company’s current annualized revenue has already exceeded $50 million, while in October last year the figure was still just above $10 million.

From a valuation perspective, this round of financing corresponds to an approximately 24x forward revenue multiple. In today’s AI startup funding market, that is relatively low.

In a December report by The Information, it said the average forward revenue multiple for AI application companies is about 53x. This means OpenRouter has maintained a degree of conservatism in its valuation—or leaves room for later investors.

The report also points out that OpenRouter is not without competitive pressure.

• At the level of large technology companies, major cloud service providers such as Amazon, Microsoft, and Google all offer multi-model access capabilities, but these platforms often tend to prioritize promoting their own models, leaving them with inherent limitations in neutrality.
• At the startup level, PortKey, as well as Vercel—which is known for hosting websites and AI applications—also offer model routing services similar to OpenRouter.
• In addition, LMArena and OpenRouter have also both released performance-based AI model rankings, creating a certain degree of indirect competition.

OpenRouter’s differentiated advantage lies in its platform neutrality and the breadth of its model coverage—access to more than 300 models, as well as a product positioning focused on developers’ workflows—which helps it maintain a first-mover advantage in the niche track of multi-model orchestration.

Risk warning and disclaimer

        There are risks in the market; invest with caution. This article does not constitute personal investment advice, and it has not considered the specific investment objectives, financial situations, or needs of individual users. Users should consider whether any opinions, viewpoints, or conclusions in this article align with their specific circumstances. Invest at your own risk, and assume responsibility accordingly.