In an modern transfer, ElevenLabs has built-in the DeepSeek R1 mannequin with its Conversational AI platform, offering a voice to the open-source mannequin. This collaboration goals to reinforce consumer interplay with AI, based on ElevenLabs.
DeepSeek R1 and Its Capabilities
The DeepSeek-R1-Distill-Qwen-32B mannequin, a distilled model of DeepSeek-R1 primarily based on Qwen2.5, has been employed for this integration. This mannequin is famous for outperforming OpenAI-o1-mini throughout varied benchmarks, setting new requirements for dense fashions. The selection of the distilled model is strategic, because the pure variations of DeepSeek R1 usually are not but dependable for operate calling, which is essential for seamless AI interactions.
Using Customized LLM Choices
The mixing leverages ElevenLabs’ Customized LLM choices, that are appropriate with any OpenAI-compatible API supplier that helps software use or operate calling. The present implementation makes use of Cloudflare, though guides for different platforms like GroqCloud and Collectively AI can be found.
Getting Began with ElevenLabs Conversational AI
To start utilizing this integration, customers can create a brand new AI agent on the ElevenLabs Conversational AI app web page. The Math Tutor template is beneficial for showcasing DeepSeek’s reasoning capabilities. This template features a system immediate that converts mathematical expressions into pure language, enhancing the interplay expertise.
For these serious about experiencing the capabilities of DeepSeek, a demo is accessible on ElevenLabs’ platform, permitting customers to work together with the AI agent and discover its functionalities.
Future Developments
Whereas the present model of DeepSeek R1 doesn’t totally help operate calling, ongoing developments intention to deal with these limitations. Customers can comply with the progress of those enhancements by way of updates supplied by the builders.
Picture supply: Shutterstock