Errors in medicine are a significant problem, highlighted as a global safety priority. General Practice is one clinical arena where error is more likely due to clinical decisions being made on a background of clinical complexity, undifferentiated symptoms and diseases, and multiple other factors as yet unquantified. Interventions designed to reduce error are either underutilised, untested, fail to produce lasting results, are designed on inadequate knowledge, or have failed to appreciate the interaction of multiple factors, both cognitive and systemic. We present a potential solution, in the form of GP Benchmark. GP Benchmark is an online simulation environment and tool designed to test clinical decision making in a group of practicing General Practitioners. Its aim is to address two pressing requirements: 1) the need to capture clinical decision making in real-time, in the context of personality, cognitive bias and environmental factors, and 2) the need to provide a validated platform that models the clinical environment so future intervention decisions may be tested without risking patient safety. We highlight the requirements satisfied for implementing GP Benchmark, the plans for validation, and discuss how GP Benchmark will be used to identify further requirements necessary to develop the environment into a tool for testing clinical decision support systems and error prevention strategies.