Pardeep Singh

🚨Big Announcement🚨 RunPod pricing is dropping by up to -40% on Serverless and up to -18% on Secure Cloud. Why We're Doing This GPUs aren't cheap, nor is the infrastructure to run them at scale. But we believe that great ideas shouldn't be held back by budget constraints. We've been fortunate to secure some serious funding recently. And instead of blowing it all on fancy office chairs or an in-house barista (tempting as that was), we chose to invest the savings in you. 1. Infrastructure Optimization: We've streamlined our operations, allowing us to pass savings directly to you. 2. Enhanced Support: We're focusing our support team to be closer to you, our customers. This means faster response times and more personalized assistance. 3. Platform Improvements: We're continuously working on reducing cold start times, enhancing our API, and introducing new features to make your experience smoother. By optimizing our pricing, we're not just cutting costs – we're reinvesting in the platform and community to provide you with a better overall experience. So let's cut to the chase. Here's what's changing: Secure Cloud We've trimmed Secure Cloud prices across the board, with some models seeing cuts of up to 18%. Serverless That's right, we've slashed prices by up to 40% on some models. Your AI workloads just got a whole lot more affordable. If you want to dive deeper into the numbers, you can check out our complete pricing page for a full breakdown of all our GPU options: runpod.io/pricing

Excited to announce our $20 million Seed led by @intelcapital and @DellTechCapital to make AI training and inference as seamless as possible at scale. A huge thank you too our investors, early supporters, and all the devs and companies that use and love our platform ❤️ We're just getting started🚀

There's been a lot of hype around AMD's MI300X recently, but very few benchmarks have been made open-source. So, we ran a series of benchmarks on Mixtral 7x8B and compared AMD's performance against Nvidia's H100 SXM. TL;DR: - MI300X outperforms H100 SXM at both small and large batch sizes (1, 2, 4, and 256, 512, 1024). - MI300X has lower time-to-first-token latency, though the difference is marginal (36.89ms vs. 55.24ms). - MI300X is generally more cost-effective than H100 SXM at small and large batch sizes. Our blog post dives deeper into the data, detailing our methods and providing code snippets so you can run the benchmarks yourself. Overall, it seems like the MI300X could rival the H100 SXM for running larger models like Mixtral 7x8B that don't fit on a single H100. But we're not ready to make definitive claims yet. We'll be running more benchmarks in the coming weeks to validate this. If you have a model benchmark request, let us know! blog.runpod.io/amd-mi300x-vs…

We hosted an AI startup 2v2 basketball tourney 2 days ago in SF Congrats to @_alexshaq and the team at @LeapAI_ for taking home $10k And thanks to all the homies that came out from @brevdev @WisprAI @LexicaArt @PrimeIntellect and more Comment if you're down for the next one