Fish routinely accelerate during locomotor maneuvers, and yet little is known about the dynamics of acceleration performance. In order to better understand how fish accelerate, we used a robotic model inspired by tuna to generate linear accelerations of various magnitudes and then quantified body kinematics, surface pressures, thrust forces and work along the body length. The Tunabot Flex platform is propelled by a 12V DC motor, measures 25.5 cm in total length, has yellowfin tuna-like body and tail profiles, and contains three body joints in addition to a fourth peduncle joint that allows bending. Linear accelerations of various magnitudes in which the tunabot moved rapidly forward were initiated in still water from a stretched-straight body position at zero initial velocity. Particle image velocimetry with high-speed video at 2000 fps was used to quantify body kinematics and fluid flow patterns, and we measured electrical power consumption throughout the maneuver. Linear accelerations varied from 0.7 m/s² to 3.5 m/s² with body kinematics similar to previous results from accelerating fishes. Peak accelerations resulted in a strong thrust wake with maximum velocities over 1 m/s. Electrical power consumption peaked within 10 ms of initiating acceleration while total mechanical power and resulting thrust force reached their initial maxima at 100 ms. The head region generated net drag, while thrust was provided by posterior body segments and the tail. Studying fish acceleration performance in an experimental platform where electrical power input, body kinematics, flow visualization, and power output into the fluid can all be simultaneously measured provides a new opportunity to understand unsteady locomotor behaviors in both fishes and bio-inspired aquatic robotic systems.