Product class
Original FMG research
A public generator dataset—not another opinion list.
We normalized 740 model records so shoppers, journalists, researchers, and AI systems can inspect the sample, reproduce the arithmetic, and see where the data is incomplete.
76%
records with a numeric price
84%
records with published weight
60%
engine models with sound data
98%
battery records with capacity
Catalog composition
What this research sample contains.
Counts below describe the FMG catalog—not universal market share. Every percentage uses the visible 740-record denominator.
Power source
Original class benchmarks
Medians and disclosure gaps by product class.
These are calculated from the normalized fields in this catalog snapshot. A dash means the measure is not meaningful or sufficiently represented for that class.
| Class | Models | Median running | Median peak | Median weight | Median listed price | Price coverage |
|---|---|---|---|---|---|---|
| Inverter generator | 325 | 3,500W | 4,300W | 85 lb | $829 | 66% |
| Portable generator | 284 | 6,250W | 7,850W | 189 lb | $999 | 78% |
| Power station | 129 | 1,800W | 3,200W | 28 lb | $699 | 96% |
| Expansion battery | 2 | — | — | 27 lb | $500 | 50% |
What the sample shows
Findings that come from FMG’s own normalization work.
Inverter generator
325 records make this the largest class in the FMG sample (44%).
4,000W
Median published running output among fuel-burning generator records with a usable rating.
60% sound coverage
368 of 609 engine-generator records include a published sound value. Missing values are not estimated.
Champion
100 model records make this the most represented brand in the current sample. This is catalog coverage, not market share.
Brand coverage
See the sample—not just the conclusion.
Brand representation reflects how many exact model records are currently normalized. A larger count does not imply higher sales, quality, reliability, or editorial preference.
Reproduce the analysis
Inspect the underlying records.
The JSON and CSV snapshots contain the normalized comparison fields, source URL, record date, and canonical product page for every model in this analysis.
Methodology
Field definitions, source hierarchy, freshness, scoring, and missing-data rules.
Evidence standard
How facts, calculations, judgments, gaps, and corrections are labeled.
Hands-on ledger
Physical testing remains separate and requires a complete public test record.
Product ledgers
Every product page exposes calculations, sources, gaps, recalls, and review dates.
Limits that travel with every finding
The catalog is curated rather than exhaustive. Manufacturer portfolios, availability, specifications, and prices change. Published sound values may use different distances, loads, modes, and test conditions. Medians describe only records with usable values, and an observed price is not a transaction price or delivered cost. No aggregate statistic proves safety, reliability, longevity, or suitability for a specific load.