Training personnel to work with animals presents a variety of challenges, both logistically and with regard to animal welfare. These issues make training an ideal opportunity to evaluate practices and to implement the 3R principles (refinement, replacement, and reduction). Cardiac blood collection from mice is a procedure that can compromise the 3Rs by requiring repeated practice and animal euthanasia. The development of a non-animal training model would promote the 3R principles. Our goals for the development of a new training model for cardiac blood collection from mice were to reduce the number of mice needed to achieve competency, improve our culture of care, and refine the training approach by improving competency. The training model was developed using commonly available materials. The total cost of the model was less than $15 USD per model. Two training curricula were conducted concurrently over a 5-mo period: 1) a curriculum in which trainees used the model before progressing to live mice and 2) the traditional curriculum, which used euthanized mice throughout. The measured variables included the total number of mice used, proportions of trainees who reached competency, the time needed to reach competency, method comprehension, quality of skill performance, trainer and trainee feedback, and training costs. The alternative group used at least 10 fewer mice per technician as compared with the traditionally trained group. The alternative group had a higher competency rate, with 82% (9 of 11 trainees) reaching competency compared with 60% (3 of 5 trainees) in the traditional group. Skill comprehension and quality were superior in the alternative group, as evidenced by fewer gross lesions at necropsy. Overall, personnel in the alternative group provided positive feedback with regard to the use of fewer mice, acquisition of both skill and confidence, and benefits for compassion fatigue. The use of this model is now our standard approach for training personnel in cardiac blood collection in mice. Our results demonstrate that the use of models in training curricula can enhance skill development and reduce the use of mice.