Information on air temperature and precipitation, often necessary for runoff modeling and forecasting, cannot always be obtained from ground-based observations. Therefore, the problem of selecting a source of climate data that is continuous in space and time becomes urgent. This paper evaluates the accuracy of 7 air temperature and 13 precipitation datasets for the Lena River basin in comparison with ground-based observations. For time series of average annual and seasonal air temperature and precipitation from 43 weather stations and climate products under consideration, the Pearson correlation coefficients, Root Mean Squared Error (RMSE), Kling–Gupta efficiency (KGE’) were calculated for two time periods: 1950–2024 and 1980–2024 for air temperature, 1966–2020 and 1982–2020 for precipitation. Correlation coefficients (average for the stations under consideration) with annual and seasonal mean air temperatures for 1980–2024, calculated from meteorological station observations, were found to range from 0.75 to 0.98 for all the products, while average RMSE ranged from 0.6 (GHCN CAMS) to 1.1 °C (NCEP-NCAR and NCEP/DOE Reanalysis) for annual mean temperatures. Climate products such as GHCN CAMS, JRA-3Q, ERA5-Land performed better than the others in reproducing air temperatures for both time periods. In the case of precipitation products, for the period 1982–2020 average correlation coefficients with ground-based observations ranged from 0.49 to 1.00, RMSE of annual precipitation totals was from 7 to 203 mm (the best results shown by GPCC Full Data Monthly Product, the worst — by NCEP-NCAR and NCEP/DOE reanalysis). For the period 1982–2020 the GPCC Full Data Monthly Product (KGE’ for annual precipitation = 0.97), GPCC Monitoring Product (KGE’ 0.84), CPC-Global (KGE’ 0.83) datasets showed the best agreement with the ground-based observations. For the period 1966–2020, the best performance was shown by the GPCC Full Data Monthly Product, PREC/L, and data downscaled from CRU TS4.09 with WorldClim 2.1.