The article presents an improved 8-km normalized difference vegetation index (NDVI) dataset, NDVI3g, derived from Advanced Very High Resolution Radiometer (AVHRR) instruments, extending from 1981 to the present. The dataset is composed of data from two AVHRR instruments: AVHRR/2 (1981-2000) and AVHRR/3 (2000-present). The main challenge in processing AVHRR NDVI data is the proper calibration of the instruments, particularly the dual gain introduced in 2000 on the AVHRR/3 instruments. To address this, the authors use Bayesian methods and high-quality, well-calibrated Sea-Viewing Wide Field-of-view Sensor (SeaWiFS) NDVI data to derive calibration parameters. The resulting NDVI values have an error of ±0.005 NDVI units, making it a non-stationary climate dataset suitable for studying seasonal and inter-annual variability in photosynthetic capacity. The article details the processing steps, Bayesian analysis, and evaluation of the dataset's quality, including spatial and temporal coherence, seasonal and inter-annual variability, and uncertainty. The NDVI3g dataset is available through the NASA Earth Exchange (NEX) platform for scientific collaboration and data sharing.The article presents an improved 8-km normalized difference vegetation index (NDVI) dataset, NDVI3g, derived from Advanced Very High Resolution Radiometer (AVHRR) instruments, extending from 1981 to the present. The dataset is composed of data from two AVHRR instruments: AVHRR/2 (1981-2000) and AVHRR/3 (2000-present). The main challenge in processing AVHRR NDVI data is the proper calibration of the instruments, particularly the dual gain introduced in 2000 on the AVHRR/3 instruments. To address this, the authors use Bayesian methods and high-quality, well-calibrated Sea-Viewing Wide Field-of-view Sensor (SeaWiFS) NDVI data to derive calibration parameters. The resulting NDVI values have an error of ±0.005 NDVI units, making it a non-stationary climate dataset suitable for studying seasonal and inter-annual variability in photosynthetic capacity. The article details the processing steps, Bayesian analysis, and evaluation of the dataset's quality, including spatial and temporal coherence, seasonal and inter-annual variability, and uncertainty. The NDVI3g dataset is available through the NASA Earth Exchange (NEX) platform for scientific collaboration and data sharing.