Today, we're thrilled to announce Together Chat, a next-generation consumer app designed to let you interact seamlessly with today's most popular open-source models, including free access to DeepSeek R1, securely hosted in the North America. Whether you're brainstorming ideas, drafting code, doing research with the web, generating images, or exploring creative writing, Together Chat puts cutting-edge open-source AI at your fingertips.

Here’s a list of some of the models that we support:

• DeepSeek R1 with web access for difficult questions
• Llama 3.3 70B / Qwen 2.5 72B for intermediate questions
• Qwen Coder 32B for coding questions
• Flux Schnell for image generation

Chat with DeepSeek R1

DeepSeek’s latest R1 model comes equipped with advanced reasoning capabilities.

Reasoning models can provide nuanced and thoughtful responses to user queries, making them a perfect fit for complex questions that current-gen LLMs struggle with.

Link: View a chat where DeepSeek reasons about how to value a SaaS company.

Search the Internet

Together Chat can even crawl the web and display search results right inside the interface.

If you toggle Web Search – or it decides your question is best answered with external sources – Together Chat will list and summarize the results, and cite its sources in its answer.

Link: See how DeepSeek summarizes real-time web search results.

Generate images

Together Chat can also generate images using Black Forest Lab’s Flux models.

Just ask for any image or drawing using natural language, and Together Chat will intelligently respond with a generated image on-the-fly:

Link: View a chat where Flux generates an image of an Eastern Bluebird.

You can also refine an image with related follow-up questions. Use image refinement to tweak your images until they’re just right.

Link: View a chat with image refinements.

Write code with Qwen

Together Chat also supports Qwen Coder 32B for generating code – an excellent coding model. You can use it for new code generation (React components or Python games for example) or to get debugging help on your existing app.

Link: View a chat where Qwen writes a React Clock component.

Lightning fast answers with Qwen and Llama

Reasoning models are great for sophisticated questions, but sometimes you just want a quick answer to a simple question.

Together Chat offers smaller open-source models like Llama 3.1 8B that respond in under a second.

Use Together Chat as a mobile app

Together Chat is a Progressive Web App, meaning it can be used as a native app on your phone. All you need to do add it to your Home Screen from the Share menu:

Now you can use Together Chat right alongside the other apps on your phone.

Start chatting with the world’s top open-source models

Together Chat is live starting today, and completely free to use.

Accounts get 10 DeepSeek R1 credits and 50 credits for all other models per day for free. Once you’ve used your daily credits, come back the following day to keep chatting.

Together Chat is built using Together AI APIs. Click here to learn more about how you can use our APIs to start powering LLM features in your own apps.