You can use OpenAI, Claude and Gemini models in one app through NanthAI Edge. Select up to three models and send them the same question in one conversation. You do not need a separate tab or account for each company to do that.
Edge reaches these models through OpenRouter, a service that sells access to hundreds of AI models through one account. The Edge site lists 300+ models that way, including models from OpenAI, Anthropic and Google, along with models from Meta and Mistral. You connect your OpenRouter account once, and every model in that list is available from the same picker.
How do you connect the models?
- Create an OpenRouter account at openrouter.ai and add some credit. The Edge site notes you can start with as little as $5.
- Open Edge, go to Settings and tap Connect OpenRouter. A standard sign-in screen links the two accounts. You do not paste an API key by hand.
- Start a new chat.
Model usage is billed by OpenRouter at the rate it lists for each model. NanthAI adds no markup on top of that, according to the Edge site. Your OpenRouter balance is shown in the chat header, and you get alerts when it runs low. For the full picture of how the costs work, see Bring Your Own Key.
How do you pick an OpenAI, Claude or Gemini model?
Tap the + button in the chat input bar and choose Participants. The picker searches by model name, so typing "claude" or "gemini" brings up the matching models. You can narrow the list with filters for free models, vision (image input), image generation and tools. The picker also has a "Help me choose" wizard if you are not sure which model fits your task.
Select up to three models. Your choice is saved for that chat, so you do not need to set it again next time.
How do you send one message to every model?
Type your question once and send it. Edge streams each model's answer into one shared card, stacked one under another, with each model's name beside its reply. Read them in turn and see where they agree and where they differ.
If you want to keep going with one of them, fork that answer into its own branch and carry on with just that model. The other branches stay one tap away. If one answer is weak, use retry-different to swap only that model for another one and keep the rest of the thread.
This works with web search, personas and knowledge base files too. Each model in the chat reads its own context, so you can ask the same question with search on and see how each one uses the sources. More on that in Multi-Model Chat.
What does access in Edge include?
- Edge does not give you the ChatGPT, Claude or Gemini apps themselves. You get the models, in Edge's own chat screen, with Edge's own features around them. Features that exist only in the company's own app may not be available here.
- Each model's answer is billed separately. A comparison costs the sum of the selected models’ usage; their prices and response lengths can differ. Check the per-message cost before you turn on three models for every message. See Price Transparency for how to see those costs.
- Edge is available on iOS, Android and the web. The cloud-model setup described here uses OpenRouter; it does not require a connected Mac.
Try it
Open the app at https://chat.nanthai.tech, connect OpenRouter, and send one question to an OpenAI model, a Claude model and a Gemini model in the same chat. For the model picker in more detail, see Participant Options.
