Takeaways
- We’re celebrating the first-ever LlamaCon by sharing some exciting new tools that will make it even easier to build with Llama and create custom models to fit your use cases.
- The Llama API, launching as a limited preview, combines the best features of closed models with open-source flexibility, offering easy one-click API key creation and interactive playgrounds to explore different Llama models.
- We’re also sharing new Llama Protection Tools and announcing the Llama Defenders Program, where select trusted partners can access AI-enabled tools to help them evaluate the security of their systems to safeguard against potential threats.
- We’re announcing the 10 international recipients of the second Llama Impact Grants. With over $1.5 million USD awarded, these grants support companies, startups, and universities using Llama to drive transformative change.
LlamaCon has officially kicked off. Our inaugural event brings together developers from around the world who all have one thing in common: They love building with Llama. It’s been a little more than two years since we launched Llama. During that time, we’ve surpassed one billion downloads and, most importantly, established Llama as a leader in the open source AI ecosystem. Developers, startups, governments, and enterprises are finding success with Llama by leveraging its capabilities to drive innovation, improve efficiency, and solve complex problems.
Of course, we couldn’t bring together a group of developers without also sharing some open source tools that will make it easier for them to explore Llama models, build faster, and use the latest tools to defend and protect. Here’s a look at what we’re announcing today and how you can get started with our newest releases.
Introducing Llama API in preview: The best features of closed model APIs, with the flexibility of open source
We want to make it even easier for you to quickly start building with Llama, while also giving you complete control over your models and weights without being locked into an API. Today, we’re announcing Llama API—our upcoming developer platform for Llama application development, which is available as a limited free preview.
Llama API provides easy one-click API key creation and interactive playgrounds to explore different Llama models, including the Llama 4 Scout and Llama 4 Maverick models we announced earlier this month. When you’re ready to build your application, we provide a lightweight SDK in both Python and Typescript. Llama API is also compatible with the OpenAI SDK, making it easy to convert existing applications.
We know that it’s also important to have access to tools that help you customize and run your models more efficiently. As part of this release, we’re sharing tools for fine-tuning and evaluation in our new API, where you can tune your own custom versions of our new Llama 3.3 8B model. We’re sharing this capability to help you reduce costs while also working toward increased speed and accuracy. You can generate data, train on it, and then use our evaluations suite to easily test the quality of your new model. Making evaluations more accessible and easier to run will help move from gut feelings to data, ensuring you have models that perform well to meet your needs. The security and privacy of your content and data is our top priority. We do not use your prompts or model responses to train our AI models. When you’re ready, the models you build on the Llama API are yours to take with you wherever you want to host them, and we don’t keep them locked on our servers. We’re introducing these features to select customers, with plans for a broader rollout in the coming weeks and months. This will open up new possibilities for developers to build custom models for every kind of use case.
This preview is just step one for the Llama API. We’re excited to have more people building with our API and will use their feedback to help us iterate on future versions. If you’re interested in applying for one of the early limited spots for our free preview, please fill out this form.
Fast inference with Llama API
We’re excited to announce a collaboration with Cerebras and Groq that will help empower more Llama developers to achieve faster inference speeds using the Llama API. Early experimental access to Llama 4 models powered by Cerebras and Groq are now available by request, offering a convenient way for developers to experiment and prototype use cases before scaling with their chosen vendor. By simply selecting the Cerebras or Groq model names in the API, developers can leverage this capability and enjoy a streamlined experience with all usage tracked in one location. This collaboration reinforces our commitment to fostering a diverse ecosystem that offers flexibility and choice, and we look forward to expanding partnerships with additional providers to bring even more options to build on top of Llama.
New Llama Stack integrations
We heard from developers that we need to make it easier to deploy applications using different service providers. Last year, we collaborated with partners to build Llama Stack distributions for their downstream enterprise clients. This year, we’re expanding those collaborations, including a recently announced integration of Llama Stack with NVIDIA NeMo microservices, and working closely with partners like IBM, Red Hat, Dell Technologies, and others on new integrations that will be announced soon. Alongside our partners, we envision Llama Stack as the industry standard for enterprises looking to seamlessly deploy production-grade turnkey AI solutions.
New Llama Protections and security for the open source community
We’re committed to providing developers with the best tools and resources to build secure AI applications. Today, we’re releasing new Llama protection tools for the open source community, including Llama Guard 4, LlamaFirewall, and Llama Prompt Guard 2. We’re also sharing updates to help organizations evaluate the efficacy of AI systems in security operations with CyberSecEval 4, as well as announcing the Llama Defenders Program for select partners. As more capable AI models become available, we believe this is an important effort to improve the robustness of software systems. Read more about our latest open source protection tools and AI security advancements news.
Meet the Llama Impact Grant recipients
The Llama Impact Grants are aimed at fostering innovation and creating economic opportunities through open source AI. Today, we’re excited to announce the 10 international recipients of the second Llama Impact Grants. With over $1.5 million USD awarded to 10 international recipients, these grants support companies, startups, and universities using Llama to drive transformative change. This year’s recipients showcase a diverse range of solutions, including E.E.R.S. from the US, which uses a Llama-enabled chatbot to enhance public access to civic services, and Doses AI in the UK, which transforms pharmacy operations with real-time error detection. Other notable winners include Solo Tech, which provides offline AI support in rural US communities, and FoondaMate, a multilingual study tool that aids millions of students across Africa and beyond. Read more about the Llama Impact Grant recipients.
The future is open source
We’re proud to be leaders in open source AI and have our Llama ecosystem at the forefront of this revolution. With Llama, developers and enterprises have the freedom to build whatever they want, without being tied down by proprietary systems or locked-in contracts. This flexibility, combined with its portability and accessibility, makes Llama the go-to choice for those looking to harness the power of AI.
We’re committed to being a long-term partner for enterprises and developers and providing a seamless transition path from closed models. Llama is affordable, easy-to-use, and enabling more people to access the benefits of AI regardless of their technical expertise or hardware resources. We believe in the potential of AI to transform industries and improve lives, which is why we’re excited to continue supporting the growth and development of the Llama ecosystem for the benefit of all. We can’t wait to see what you’ll build next.