NVIDIA RTX 5090 outperforms AMD and Apple running local OpenAI language models

Here is a 186-word summary of the article: Developers and creatives are increasingly turning to locally run AI models like OpenAI's new GPT-OSS family, which can run on consumer GPUs with just 16GB of memory. Among the hardware options, NVIDIA's GeForce RTX 5090 GPU has emerged as the best performer for running these open-weight AI models. Tests by the Llama.cpp open-source framework show the RTX 5090 achieving 282 tokens per second on the GPT-OSS-20B model, outperforming the Mac M3 Ultra (116 tok/s) and AMD's 7900 XTX (102 tok/s). This is thanks to the RTX 5090's Tensor Cores designed to accelerate AI workloads. For easy experimentation with GPT-OSS models, the article recommends the LM Studio and Ollama applications, both of which are optimized for NVIDIA GPUs. These tools handle model downloads, environment setup, and GPU acceleration, making it simple for users to test the latest open-source language models on their local hardware.