A limited program for AI teams running real inference at scale — cut GPU costs, reduce infra work, and serve models faster.
A limited program for AI teams running real inference at scale —
cut GPU costs, reduce infra work, and serve models faster.
The FriendliAI Launch Credit Program provides eligible production stage AI teams with up to $10,000 in GPU inference usage credits, allowing you to experience our fast, cost effective, and fully managed AI inference service without any operational hassle.
FAQs
Yes—once credits are applied to your account, you can use them across multiple deployments, including different models, endpoints, or applications. Whether you're testing fine-tuned variants, scaling inference, or comparing configurations, usage is flexible within your quota.
Selected teams can receive up to $10,000 in free inference credits. The exact amount awarded is tailored to your team’s needs and current usage. Our goal is to provide a credit amount that meaningfully supports your AI workload – this could be the full $10K for teams with very high GPU usage, or a smaller credit tier for teams with lower requirements. In every case, we’ll discuss your usage to ensure you get an appropriate level of support.
The program is best suited for AI teams running or about to run production workloads. That includes services with paying customers, active free users, or high-volume internal use cases that mirror real-world usage. While we welcome a range of applicants, our goal is to award credits to teams who can actively test FriendliAI under meaningful conditions—where cost, speed, and reliability matter most.
Not necessarily. Even if you're in a limited beta or have a test user group, you may still qualify—especially if you're expecting real traffic soon. The key is that you have an actionable plan to use the credits in a realistic setting. We’re looking for teams who can quickly activate and make full use of the program within the 30-day credit window.
The program has a quick, three-step selection process. We aim to complete everything in about 10 days from your application submission:
(Note: We strive to notify all applicants of the outcome. If you’re not selected for a discovery call or the program, we will inform you as well in a timely manner.)
The discovery call is a 30-minute, collaborative conversation with our team. Since this is a selective program, we use the call to better understand your project’s goals, model deployment plans, and real-world workload—like expected usage scale and infrastructure needs. We’ll also walk through how Friendli might support you, tailor the credit tier accordingly, and leave time for any questions you have. It’s a practical step to make sure there’s a strong fit on both sides.
If your team is accepted, we will apply the credits directly to your FriendliAI account for you to use. If you don’t have a FriendliAI account yet, don’t worry – we’ll help you set one up as part of the onboarding. Once the credits are applied, any usage of FriendliAI’s inference services (deploying models, running API calls, etc.) will automatically deduct from your credit balance. You’ll be able to start deploying models and serving inference workloads immediately, without incurring charges until the credit is used up. Our team will provide guidance to get you up and running quickly so you can make the most of your credits.
No – there are no contracts to sign and no long-term commitments required to use the free credits. This program is meant to let you try FriendliAI’s inference platform risk-free. You won’t be locked into any subscription or obligation beyond our standard terms of service. After using the credits, it’s entirely up to you whether you continue with a paid plan. Of course, we’re confident you’ll see value in the platform (in terms of cost savings and performance), but continuing to use FriendliAI after the credits is your choice – not a requirement.
The credits are designed to support your real-world usage, so their validity period begins once you activate them. From that point, you’ll have 30 days to use the credits—plenty of time to evaluate how FriendliAI improves inference cost and performance. While there’s no rush to activate right away, we recommend starting when your project is ready to move forward. If you have concerns about the time frame or need a bit longer to ramp up, just let us know—we’re flexible and happy to work with you to ensure you get the full value from your credits.
After your credit is used up, there’s no obligation to continue – the free credit involves no strings attached. If you decide not to move forward with FriendliAI, you can simply stop using the service at that point. However, if you’ve found value in our platform (many teams continue because they see significant speed-ups and cost savings), you have the option to continue as a regular customer. This would mean transitioning to our standard usage-based billing or a suitable plan going forward. We’ll be happy to discuss next steps with you when the time comes – whether that’s setting up a paid plan that fits your budget, or gracefully concluding the trial. The bottom line is that the choice is yours, and we’ll support you either way.
FriendliAI is a GPU platform for accelerated AI, built to make serving AI models faster, more efficient, and easier to scale. Integrated with Weights & Biases & Hugging Face, FriendliAI enables instant model deployment, traffic-based autoscaling and significant GPU cost savings so you can deliver reliable inference without managing infrastructure.
TECH BLOG
Hugging Face
Models
Weights & Biases
W&B
AI DevOps
Multimodal
Inference
Optimization