We're excited to introduce web search capabilities for all AI models on PayPerQ. Whether you're using the latest frontier models or specialized alternatives, you can now enhance your conversations with real-time information from the web.

How to Enable Web Search

Enabling web search is simple. Just toggle the web search option in your chat interface before sending your message:

Once enabled, your AI model will have access to current information from the web to provide more accurate and up-to-date responses.

How It Works

Understanding how web search works helps you get the most out of this feature. Here's what happens behind the scenes:

1. Search First: Before your query reaches the AI model, a web search is performed based on your question
2. Results Retrieved: The search engine pulls back approximately 5 relevant results from the web
3. Context Injection: These search results are automatically injected into the prompt
4. Enhanced Response: The AI model receives your question along with the current web context, allowing it to provide informed, up-to-date answers

This process happens seamlessly in the background, giving you responses grounded in real-time information without any extra steps.

See It In Action

Here's an example of web search helping answer a question about the NYC mayoral election. In the first attempt (left), Grok 4 didn't have the election results in its training data. After toggling web search on and asking the same question again (right), the model was able to provide current, accurate information:

This demonstrates how web search bridges the gap between a model's training cutoff date and real-time information.

Two Types of Web Search

PayPerQ uses two different search engines depending on which AI model you're using:

Native Search

Some AI providers have built their own web search capabilities directly into their models. When you use these models with web search enabled, we utilize their native search functionality:

• Perplexity models (all variants)
• OpenAI models: GPT-5 and its variants
• Anthropic models: Claude 4.5 and other Claude models

Native search is optimized by each provider for their specific models and may offer unique features or enhanced integration.

Exa.AI Search

For all other models that don't have native web search capabilities, we use Exa.AI, a powerful search engine designed specifically for AI applications. Exa.AI uses a combination of keyword search and embeddings-based search to find the most relevant results for your query.

This means you can add web search capabilities to hundreds of models that wouldn't otherwise have access to real-time information.

Pricing

Web search does add a small cost to your queries, but we've kept it transparent and affordable:

Exa.AI Search Pricing

When using Exa.AI search (for models without native search):

• $0.02 per request (for 5 search results)
• This cost is in addition to the model's standard token pricing

Native Search Pricing

When using models with native search capabilities, pricing is determined by the model provider and varies based on the specific model and search context needed. Native search is generally more expensive than Exa.AI search, but the exact cost depends on factors like:

• The specific model being used
• The amount of search context required for your query
• The provider's pricing structure

While native search costs more, it offers optimized integration with each provider's models and may provide enhanced search capabilities specific to that provider.

Get Started Today

Web search is now available for all users on PayPerQ. Simply toggle it on in your next conversation to start getting responses enhanced with real-time web information.

Whether you're researching current events, looking up recent developments in your field, or need the latest data for your projects, web search ensures your AI assistant has access to the information it needs to help you effectively.

Try it out and experience the difference that real-time web access makes in your AI conversations!