This data set contains Calibrated data taken by the New Horizons Long Range Reconnaissance Imager instrument during the Pluto encounter mission phase. This is VERSION 3.0 of this data set. This data set contains LORRI observations taken during the the Approach (Jan-Jul, 2015), Encounter, Departure, and Transition mission sub-phases, including flyby observations taken on 14 July, 2015, and departure and calibration data through late October, 2016. Departure observations include a ring search of the Pluto system and 1994 JR1 observations. This data set completes the Pluto mission phase deliveries for LORRI. This is version 3.0 of this dataset. Changes since version 2.0 include the addition of data downlinked between the end of January, 2016 and the end of October, 2016, completing the delivery of all data covering the Pluto Encounter and subsequent Calibration Campaign. It includes multi- map observations from the Approach phase, observations of the moons, hi- res, full-frame observations from Pluto Encounter and Departure, sliver maps, and ring search observations. There may be some overlap between prior datasets and this dataset, due to only partial, windowed, or lossy data in prior datasets. Observations at closest approach to Pluto are marked with _CA in the Request ID. This dataset also includes functional tests from the Calibration Campaign, including a regular observation of NGC3532. Finally it includes the first set of distant KBO observations. Also, updates were made to the calibration files, documentation, and catalog files. There were minor changes to the level 2 LORRI calibration process, as well as to the LORRI calibration constants for the final Pluto P3 PDS delivery. The process change involves gap removal during calibration. Files with gaps come in many flavors, depending on where the gap lies within the image. This update recognizes some additional possiblities, mainly that the gap might be close to the bottom or top of the image (and therefore the previous algorithm would fail because it filled the gap with median pixel info from both above and below the gap). The new algorithm will take the info from one side of the gap exclusively, when appropriate.