In this Build With Nebius episode, Sujee Maniyam, Developer Advocate at Nebius, puts multiple AI models through the same real-world data-analysis task. We’ll compare accuracy, speed, token usage, tool calls, and total cost, then look under the hood at how these agents reason, use tools, and manage context. Sujee will also share an agent framework bug he uncovered, how he diagnosed it, and the fix.
Then, Jacquie Capur, Senior Developer Advocate at Nebius, will use Hermes and Nebius Token Factory to celebrate her dog’s second birthday by turning his picture into a video game. Join live to vote for the levels you want to see built as we explore different model types, model judge loops, and evaluations. Can we build a cute 2D platformer in 30 minutes? Let’s find out!
Resources
Images
