DuckDuckGo adds access to GPT-5, Claude 4, and other AI models to its subscription plan
DuckDuckGo has expanded its paid subscription, formerly known as Privacy Pro, to include advanced artificial intelligence models from top providers such as OpenAI, Anthropic, and Meta. Through the Duck.ai platform, subscribers can now utilize AI tools including OpenAI’s GPT-4o and GPT-5, Anthropic’s Claude Sonnet 4, and Meta’s Llama Maverick.
The subscription price remains the same. Users continue to pay $9.99 per month or $99 per year. In addition to new AI offerings, the subscription still provides access to DuckDuckGo's virtual private network (VPN), Personal Information Removal, and Identity Theft Restoration services.
While introducing these upgrades, DuckDuckGo has indicated plans for higher subscription tiers in the future. These forthcoming options are expected to grant access to even more powerful AI models.
Users who prefer a non-AI experience retain control over their settings. Duck.ai features remain optional for all, regardless of subscription status. The AI buttons and related features can also be hidden directly from search settings, as well as on desktop and mobile browsers.
- AI Chatbot
- Freemium • Proprietary
Related news
External links
- DuckDuckGo subscription adds advanced AI chat modelsDuckDuckGo • Official source
- DuckDuckGo adds access to advanced AI models to its subscription planTechCrunch
- DuckDuckGo’s subscription now includes access to GPT-5 and GPT-4oThe Verge
- DuckDuckGo introduces paid tier with advanced OpenAI & Anthropic modelsAndroid Headlines
- DuckDuckGo's subscription now offers some of the latest chatbots from OpenAI and AnthropicEngadget
Comments
Not clear what the context length and rate limits are. I would get it if they had true ZDR.
DDG states clearly that it doesn't want to give details on usage limits to prevent users from abuse them, most users being on free plan. Since DDG doesn't save the IP, users could easily switch browsers or configurations to create many chatbots, so it must use some secret and changing formula to throttle users in some way.
I mean on the paid $10 plan.
I feel like context length is usually a bit smaller than what you need for that convo lol. Gemini does great with the 1M + token length but the openai models often struggle with long convos, forgetting things here and there due to the constraints.