MENLO PARK — Together AI, a leader in AI acceleration cloud services, has taken the developer world by storm with the release of LlamaCoder, an open-source web app powered by the Llama 3.1 405B model. This innovative tool allows developers to generate fully functional apps from a simple text prompt, showcasing the power and flexibility of Llama’s AI capabilities.

Since its release, LlamaCoder has become a viral sensation. In just over a month, it has garnered more than 2,000 stars on GitHub, with hundreds of developers cloning the repository and over 200,000 apps generated using the platform.

“Developers are really embracing it,” says Hassan El Mghari, the software engineer leading developer relations at Together AI. “Seeing people create everything from quiz apps and budgeting tools to productivity timers simply by using prompts highlights the immense potential of Llama 3.1 405B.”

Llama 3.1 405B is the first open-source model that excels in generating code for such specific use cases. Together AI has also developed additional example apps, such as LlamaTutor, which assists users in learning new skills, and TurboSeek, an AI-driven search engine.

“Llama models are redefining the boundaries of open-source large language models (LLMs), rivaling some of the best closed-source alternatives, while also providing essential safety tools for responsible AI development,” explains El Mghari. “The capabilities of Llama 3.1 405B in areas like general knowledge, multilingual translation, and tool use have already proven to be a game changer for both developers and enterprises.”

Together AI’s platform is now being utilized by over 150,000 developers and companies for a wide variety of applications, ranging from gaming and customer service to summarization and benchmarking. The platform allows users to retain ownership of their fine-tuned models and easily apply them to new use cases, enhancing the flexibility and scalability of AI development.

“Our mission is to quickly bring cutting-edge research innovations to production,” says El Mghari. “Built on our advanced inference optimization research, the Together Inference Engine leverages state-of-the-art innovations like FlashAttention-3 kernels and RedPajama-based speculators, which enable unmatched performance, accuracy, and cost-efficiency at production scale.”

A key factor behind this success is Together AI’s commitment to open-source development. According to El Mghari, many companies have transitioned to Together AI from closed-source providers, recognizing the benefits of open-source AI, such as increased flexibility and ownership of their data and models.

“Open-source AI has revolutionized the industry,” says El Mghari. “It’s driving faster innovation and greater collaboration within the developer community, all while giving organizations complete control over their models and data, rather than locking them into closed ecosystems.”