SAN FRANCISCO — Together AI is proud to collaborate with Meta in advancing the Llama model series with the introduction of Llama-3.3-70B-Instruct. This latest addition sets a new benchmark in artificial intelligence, delivering exceptional performance in reasoning, mathematics, and instruction-following tasks. By leveraging state-of-the-art post-training techniques, Llama-3.3-70B achieves capabilities comparable to the larger Llama 3.1 405B model, all while being more cost-efficient and accessible.

Our platform empowers over 250,000 developers and enterprises, including renowned names like Zomato, Salesforce, DuckDuckGo, and The Washington Post. Together AI is enabling these innovators to harness AI’s transformative potential, driving impactful customer experiences and cutting-edge solutions.

Revolutionizing AI Deployment with Serverless Endpoints

To fully support Llama-3.3-70B, we’re introducing Together Turbo serverless endpoints. These endpoints redefine AI deployment by combining unmatched speed, accuracy, and affordability. With FP8 technology at its core, the Llama-3.3-70B endpoint offers lightning-fast computations without compromising model quality, ensuring seamless integration for developers and businesses alike.

Dedicated Endpoints for Consistent and Reliable Performance

For those requiring dependable, high-capacity solutions, we will soon roll out Dedicated Endpoints for Llama-3.3-70B. These endpoints guarantee consistent performance by eliminating shared resource constraints, making them ideal for critical applications. With no rate limits and robust load capacity, Dedicated Endpoints deliver the reliability needed for projects demanding low latency and uninterrupted service, whether scaling large applications or crafting specialized tools.

Championing Open-Source AI Innovation

At Together AI, we’re committed to fostering an open-source ecosystem that prioritizes transparency, adaptability, and collaboration. Llama-3.3-70B is available under a community license, allowing developers to customize and innovate freely. By embracing open-source principles, we help mitigate vendor lock-in, giving developers full control over their AI infrastructure and data.

This model exemplifies the strength of open-source AI, empowering businesses, researchers, and developers to push boundaries and create impactful solutions. Together AI’s dedication to this vision ensures that powerful, accessible AI technology is within reach for everyone.

Start Building Today

Learn more about how Together AI and the Llama 3.3-70B

Reach out to our sales team to discover how Llama-3.3-70B and Together AI can help you achieve your goals.