In this work, we evaluate the SUPIM-INPE model prediction of the 14 December 2020, total solar eclipse over the South American continent. We compare the predictions with data from multiple instruments for monitoring the ionosphere and with different obscuration percentages (i.e., Jicamarca, 12.0°S, 76.8°W, 17%; Tucumán 26.9°S, 65.4° W, 49%; Chillán 36.6°S, 72.0°W; and Bahía Blanca, 38.7°S, 62.3°W, reach 95% obscuration) due to the eclipse. The analysis is done under total eclipse conditions and non-total eclipse conditions. Results obtained suggest that the model was able to reproduce with high accuracy both the daily variation and the eclipse impacts of E and F1 layers in the majority of the stations evaluated (except in Jicamarca station). The comparison at the F2 layer indicates small differences (<7.8%) between the predictions and observations at all stations during the eclipse periods. Additionally, statistical metrics reinforce the conclusion of a good performance of the model. Predicted and calibrated Total Electron Content (TEC, using 3 different techniques) are also compared. Results show that, although none of the selected TEC calibration methods have a good agreement with the SUPIM-INPE prediction, they exhibit similar trends in most of the cases. We also analyze data from the Jicamarca Incoherent Scatter Radar (ISR), and Swarm-A and GOLD missions. The electron temperature changes observed in ISR and Swarm-A are underestimated by the prediction. Also, important changes in the O/N2 ratio due to the eclipse, have been observed with GOLD mission data. Thus, future versions of the SUPIM-INPE model for eclipse conditions should consider effects on thermospheric winds and changes in composition, specifically in the O/N2 ratio.