Hackathon Project OCCAM Measures Whether Agentic Steps Are Worth Their Tokens

On October 3, 2026, DEV Community published a post in its AI section under the headline "Is your AI agent worth its tokens? We measured it with TigerGraph." The author describes work carried out during the TigerGraph Agentic GraphRAG Hackathon, where the team built a project called OCCAM. Its aim was to work out whether the extra reasoning and retrieval steps an AI agent performs are worth the extra tokens those steps consume.

The question the author puts forward is not whether an agentic setup delivers a better answer, but whether the added steps pay off once token cost is taken into account. According to the post, the hackathon guidebook frames the problem in the same way: the real question is not whether agentic produces a better answer, but whether the extra reasoning and retrieval steps are worth the extra token cost.

The post opens with the observation that everyone is bolting agents onto retrieval, and almost nobody asks what they cost. It carries tags including ai, rag, graph and python.

That is the full extent of what the available reporting covers. OCCAM is presented as a tool built during the hackathon to measure the token cost of an AI agent's added steps, and the piece is organized around the cost question rather than around answer quality. The material contains no measurement results, no figures, and no comparison of any kind, and it does not say whether the project was developed further or released after the hackathon.