In this tutorial, you will run the DeepSeek-V4-Flash-0731 large language model across two DGX Sparks using sparkrun. This model is a Mixture-of-Experts model with 284 billion total and 13 billion active parameters. Additionally, before running it, we will examine how intelligent the model is.<\/p>\n
The Spark can be used as a standalone computer by connecting a monitor and keyboard, or as a remote server accessed from another computer. In this tutorial, we will connect to the Spark remotely and install the necessary software to run the model.<\/p>\n
We will use vLLM as the inference engine and sparkrun as the management tool. vLLM will load the model’s trained weights into GPU memory and expose an API that accepts external requests. sparkrun will manage Docker containers and model deployment across Sparks from the command line. This way, operations such as image synchronization, model transfer, and cluster configuration will be automated. Both will run on the Main Spark, inside Docker containers.<\/p>\n
This tutorial consists of six parts:<\/p>\n
- \n
- Model Intelligence:<\/strong> Tests that measure the intelligence of AI models, how they differ from performance tests, and DeepSeek-V4-Flash-0731’s position among leading models<\/li>\n
- Setup:<\/strong> Downloading the Docker image and preparing the recipe<\/li>\n
- Running:<\/strong> Starting, monitoring, and testing the model on two DGX Sparks with sparkrun<\/li>\n
- Benchmark:<\/strong> Performance measurement results at different concurrency levels<\/li>\n
- Using with OpenCode:<\/strong> Connecting the model to a local coding assistant<\/li>\n
- Shutdown:<\/strong> Stopping the services<\/li>\n<\/ul>\n
Throughout this tutorial, the primary device will be referred to as the Main Spark and the secondary device as the Worker Spark.<\/p>\n
\nModel Intelligence<\/h2>\n
1. Performance Tests vs Intelligence Tests<\/h3>\n
When “AI Benchmark” is mentioned, performance tests generally come to mind. In our previous tutorials, we primarily focused on these tests: we ran various models known to be intelligent on the Spark. Then we tested values such as how many tokens the model could produce per second (TPS), how long it took for the first token to arrive (TTFT), or token generation at different concurrency levels (throughput).<\/p>\n
Performance tests are fundamentally about how fast a model responds. They depend on which hardware and software a particular model runs on. However, they do not tell us whether the model’s responses are correct. A model can run at 70 tok\/s but produce incorrect or nonsensical responses.<\/p>\n
Intelligence tests, on the other hand, are fundamentally about the accuracy of the model’s responses. They are hardware-independent \u2014 a property related to the model’s trained weights.<\/p>\n
2. Artificial Analysis and Intelligence Indices<\/h3>\n
- Setup:<\/strong> Downloading the Docker image and preparing the recipe<\/li>\n
![](\"https:\/\/blog.openzeka.com\/en\/wp-content\/uploads\/2026\/08\/aa-intelligence-index-1024x662.webp\" "\\"aa-intelligence-index\\"")
<\/span><\/div>![](\"https:\/\/blog.openzeka.com\/en\/wp-content\/uploads\/2026\/08\/aa-coding-index-1024x662.webp\" "\\"aa-coding-index\\"")
<\/span><\/div>