creativitybenchmark.ai

What It Is

Creativity Benchmark is an AI creativity evaluation benchmark launched by Springboards.ai. It is not positioned as a typical content generation tool, but rather as a collective evaluation project for the advertising, marketing, design, and strategy industries. Its goal is to fill a gap in existing AI benchmarks, which tend to focus on math, science, law, and coding while paying less attention to creative thinking.

Core Capabilities

The platform’s basic mechanism is to let users vote between two anonymous model outputs and choose the better response, using those votes to measure model performance. The evaluation focuses on three areas: creative problem-solving, meaning whether a model can propose novel and diverse solutions; insight and idea generation, meaning whether the output can inspire stronger starting points for thinking; and variance, meaning whether a model’s outputs become repetitive or show sufficient variety across a large number of generations. The page also mentions that users can view their own model preferences.

Pricing and Data

The main page does not disclose any commercial pricing. It only states that people in the industry can participate and that the process takes just a few minutes, so it appears more like a free, research-oriented participation project. On privacy, the page clearly promises that outputs from the evaluation will not be used to train any models, which is an important plus. However, for participant data collection, usage methods, the list of models, and the research timeline, the captured content only shows FAQ headings and does not provide specific answers.

Pros and Cons

Its strengths are clear positioning, a focus on real-world tasks in creative industries, and endorsements from organizations such as 4As, D&AD, IPA, IAA, The One Club, as well as multiple industry advisors, which gives it solid credibility. The voting interaction is simple and suitable for creative professionals without a technical background. Its limitations are that the publicly available information is still more of an initiative introduction: it lacks details on specific models, evaluation methodology, statistical criteria, and published results. There is also no visible information about an API, enterprise integrations, or Chinese-language support.

Who It’s For and Access from China

It is suitable for strategists, advertising creatives, brand marketing teams, and creative technology leads who want to understand which models are better suited to creative and strategic tasks. It can also serve as one reference point for internal model selection. However, it is not a production tool that directly replaces ChatGPT, Claude, or Midjourney. There is no clear information about access, accounts, or payments from mainland China, so it is best treated as “unknown.” If access is unstable, alternatives include Chatbot Arena, HELM, Artificial Analysis, or building an internal creative evaluation workflow.