Show HN: Slopsome – a VRAM fit calculator and tok/s database for local LLMs (slopsome.com)

A new tool named Slopsome has been introduced, functioning as a VRAM fit calculator and token-per-second (tok/s) database specifically designed for local large language models (LLMs). This user-friendly platform allows users to input their GPU specifications—such as the NVIDIA RTX 4090—and immediately see which LLMs can run on their hardware, along with estimates on performance metrics like tokens per second. For instance, it informs users that the Qwen3 32B model requires 21.1 GB of VRAM and can process around 38 tokens per second, making it suitable for the RTX 4090. This release is significant for the AI/ML community, particularly as more developers seek to deploy LLMs for various applications without relying solely on cloud resources. By providing clear insights into resource requirements and performance capabilities, Slopsome democratizes access to LLM experimentation, allowing users to make informed decisions based on their hardware limitations. The interactive nature of the platform empowers developers to optimize their setups, fostering innovation in local AI implementations while avoiding the pitfalls of overestimating hardware capabilities.