Methodology v3 • October 2026
Live in a few categories, more coming soon
How the overall score works (methodology v3)
Each model gets five marks out of 10. The overall score is their weighted average, from 1 to 9.9. Each category has its own weights, the same for every model in it.
Technical score: Built from the model's specifications. Each spec is compared with the other models in the category and weighted by how much buyers care about it.
Reviews score: We read and grade independent written and video reviews from test sites, magazines and video reviewers. The mark is 80% the average grade and 20% the number of reviews. With only a few reviews, the average is pulled toward the typical grade of the category, so one review cannot carry a model to the top.
List score: How many independent best-of lists name the model. The model in the most lists gets 10, the others get 8 to 10 in proportion.
Sentiment score: What owners say after using the model, from their comments under video reviews and in forums. The mark is 80% how satisfied they are and 20% how many owners we found. It counts from 10 owners on.
User score: Star ratings that buyers give the model online, for all its versions and in several countries. The mark is 70% the average rating (5 stars give 10) and 30% how many ratings the model has (one rating gives 5, a typical number 8, the most rated models 10). With only a few ratings, the average is pulled toward the typical rating of the category. It counts only in categories where at least 10 models have ratings.
For reviews, lists and owner comments, a typical model in the category gets about 8 and the best one gets 10. The User score starts from the average star rating itself (4 stars give 8), so a model with typical ratings lands a little above 8.
What a "?" means
We do not have enough data for that part yet. The part still counts, as the mark of a weaker model in the same category. If we left it out, a model nobody has tested could beat a model with good tests.
Reviews score
Shows "?" when
no independent review yet
Counts as
the mark that 4 in 5 reviewed models in the category beat; it differs per category
Sentiment score
Shows "?" when
fewer than 10 owner opinions
Counts as
the mark that 4 in 5 models with owner data beat; it differs per category
List score
Shows "?" when
in no best-of list
Counts as
7.5 in every category, just below the 8 to 10 of listed models
User score
Shows "?" when
no ratings yet
Counts as
the mark that 4 in 5 rated models in the category beat; it differs per category
The hover window of each "?" shows the exact number used.
Each category uses its own mix of these scores, set from how much data exists for it: the more expert reviews and best-of lists a category has, the more they count. This keeps the ranking as objective and balanced as possible. Each category's methodology page shows its own mix.
Each category's methodology page shows the version it uses.
