CEC efficiency vs peak efficiency: which to use in proposals
Use the CEC weighted efficiency, not the peak. The peak is the single best point on an inverter's test curve; the CEC figure weights the curve by how often an inverter runs at each power level. For the seven inverters here the CEC figures ran from 97.0% to 99.0%, and the peaks were up to 0.7% higher. Over a modeled year in West Covina, every inverter averaged less than its peak, by at least 0.2%, and came within 0.9% of its CEC figure. The biggest gap was Powerwall 3 running a modest array: rated 97.5% by the CEC with a 98.2% peak, it averaged 96.6% on 8 kW of panels, because a large inverter spends more hours at low load.
In this guide
Three efficiencies, seven inverters#
Swipe to see all columns →
| Inverter | Rated, kW | CEC weighted | Peak on the CEC curve | Over a year in West Covina (model) | Peak minus the year |
|---|---|---|---|---|---|
| Enphase IQ8+ | 0.295 | 97.0% | 97.7% | 96.4% | 1.2% |
| Enphase IQ8M | 0.325 | 97.0% | 97.7% | 96.3% | 1.4% |
| Enphase IQ8A | 0.349 | 97.0% | 97.6% | 96.5% | 1.1% |
| Enphase IQ8HC | 0.384 | 97.0% | 97.4% | 96.8% | 0.6% |
| Enphase IQ8X | 0.384 | 97.0% | 97.7% | 96.8% | 0.9% |
| SolarEdge SE7600H | 7.616 | 99.0% | 99.0% | 98.8% | 0.2% |
| Tesla Powerwall 3 (11.5 kW) | 11.508 | 97.5% | 98.2% | 96.6% | 1.6% |
CEC weighted efficiency and rated output from the CEC Grid Support Inverter List (downloaded October 8, 2026); peak is the highest of the 18 efficiency points the CEC list gives for each inverter (six power levels at three voltages). "Over a year": the Cali-Graf engine's Sandia model of each inverter with LONGi LR5-54HABB-400M panels facing south at 20° in ZIP 91790, typical-year weather; 12 panels on the microinverters, 20 panels on one SolarEdge SE7600H or one Powerwall 3. Clipping is excluded. Source: CEC; Cali-Graf engine.
Why the peak overstates#
- Inverters rarely run at their best point. Efficiency is lower at low power, in the morning, evening and on cloudy days, and slightly lower at full power; the peak is one point on the curve. - Size against the array matters. A large inverter on a small array spends more of the year at low load: Powerwall 3, rated for up to 11.5 kW, on 8 kW of panels averaged 96.6%. - Voltage matters. The CEC curve is tested at three DC voltages; the peak is the best of all of them, while a real array runs at its own voltage.
What to put in a proposal#
- Quote the CEC weighted efficiency. It is published for every listed inverter and comes within about a point of a modeled year, which the peak does not. - For SolarEdge, add the optimizers. The CEC figure covers the SolarEdge inverter alone; the power optimizers on each panel have losses of their own, given in their data sheet, and our model does not include them, so compare a SolarEdge system with microinverters only after adding them. - Better, model the hours. A production model that runs the inverter's own curve against the site's weather, as the Cali-Graf engine does, replaces a single efficiency figure altogether. - Don't double count. A production estimate that already models the inverter should not also apply a separate inverter loss.
Data limits#
The engine's inverter parameters come from the CEC library used by the System Advisor Model, which can differ slightly from the CEC list's figures; one panel model, one site and one typical year; clipping is excluded from the yearly efficiency. The CEC list rounds most weighted efficiencies to a whole percent. The model does not include SolarEdge's optimizer losses.
What is CEC weighted efficiency?
The CEC's efficiency figure for an inverter, weighting its measured efficiency at different power levels by how often inverters run at them, published on the CEC's equipment list.
Is peak efficiency or CEC efficiency more realistic?
CEC: in our model the year's average came within 0.9% of the CEC figure for every inverter here, and was always below the peak.
Why is Powerwall 3's efficiency lower in my estimate than on the datasheet?
A large inverter on a small array runs more hours at low load; on 8 kW of panels our model averaged 96.6% against the CEC's 97.5%.
Methodology#
CEC Grid Support Inverter List, downloaded October 8, 2026: maximum continuous output, weighted efficiency and the efficiency table at 10% to 100% power and minimum, nominal and maximum DC voltage. The Cali-Graf engine ran a year of hourly output with each inverter's Sandia model (CEC parameters in the System Advisor Model library) and LONGi LR5-54HABB-400M panels from the CEC module list; yearly efficiency is one minus the inverter's losses divided by its DC input, with clipping excluded. Rounded half up.
Sources#
- California Energy Commission — Solar Equipment Lists, Grid Support Inverter List. https://solarequipment.energy.ca.gov/
- Sandia National Laboratories — Performance Model for Grid-Connected Photovoltaic Inverters (SAND2007-5036). https://www.osti.gov/biblio/920449
- NLR — PySAM, the Python package for the System Advisor Model. https://nrel-pysam.readthedocs.io/
- Tesla — Powerwall 3 Datasheet (2025). https://energylibrary.tesla.com/docs/Public/EnergyStorage/Powerwall/3/Datasheet/en-us/Powerwall-3-Datasheet.pdf
How to cite this page#
Cali-Graf. “CEC efficiency vs peak efficiency: which to use in proposals.” cali-graf.com, updated Nov 19, 2026. https://cali-graf.com/guides/cec-efficiency-vs-peak-efficiency-which-to-use-in/