Why our own score
AI glasses cannot be read off from a single number. One pair may have a brilliant display and still press on your nose after two hours; another is featherlight, but its AI does not understand Swiss German. Spec sheets do not answer these questions — real everyday use does.
The AI Eyewear Score translates precisely this everyday use into a comprehensible grade. It is deliberately category-aware: a pure audio-camera pair is not measured against a display it does not even have. And it is deliberately reproducible: every criterion has defined measurement points, so that two devices remain comparable under the same conditions.
The six criteria
Every pair of glasses is scored from 0 to 100 in six areas. The areas are weighted and combined into the overall grade.
Display & optics
Legibility in everyday use — not the spec sheet.
- Brightness & contrast outdoors
- Sharpness, colour fidelity, image stability
- Field of view & positioning
- Prescription option (Rx)
Comfort & design
Whether you forget them all day long.
- Weight & weight distribution
- Pressure points after hours
- Fit, nose pads, temples
- Build quality & everyday suitability
AI, software & updates
How clever and how well-maintained the glasses are.
- Assistant quality (incl. DE/dialect)
- App, pairing, reliability
- Latency & offline capability
- Update history & support
Battery & endurance
Real runtime, not a lab figure.
- Runtime in mixed everyday use
- Charging case / topping up on the go
- Charging time & standby
- Behaviour under load (AI, display)
Privacy & transparency
What the glasses reveal about you and others.
- Camera indicator (if a camera)
- Data processing on-device vs. cloud
- Clear settings & deletability
- Signalling effect for bystanders
Value for money
What you actually get for your money.
- Price incl. necessary accessories
- Ongoing costs (Rx, case, subscription)
- Durability & repairability
- Value compared to the category
Weighting in detail
The base weighting applies to display/AI glasses (e.g. Even G2). For glasses without a display, the weight shifts transparently from "Display & optics" to audio and camera quality — so every category can be scored fairly, instead of being measured against a feature it does not possess.
| Criterion | Display/AI glasses | Camera/audio glasses (without display) |
|---|---|---|
| Display & optics | 22 % | — (not applicable) |
| Audio & camera (only without display) | — (included in optics/AI) | 22 % |
| Comfort & design | 20 % | 20 % |
| AI, software & updates | 18 % | 18 % |
| Battery & endurance | 15 % | 15 % |
| Privacy & transparency | 15 % | 15 % |
| Value for money | 10 % | 10 % |
| Total | 100 % | 100 % |
The weighting is part of the methodology version. If we change it, we raise the version number and document the change — older grades remain traceable with their version reference.
What the grade means
How we test — the protocol
1 · Disclose sourcing
For every test we state whether the device was bought, borrowed or provided as a loan unit. Loan units are returned; provision does not change the grade.
2 · Verify specifications independently
We verify manufacturer specifications (weight, brightness, battery) ourselves — with scales, a lux meter and charge-level measurement — instead of adopting them. We record any deviations.
3 · At least two weeks of everyday use
Every pair is worn over a continuous everyday period — in the office, on the go, outdoors in daylight, in our showroom in Zurich — not just in a quick test.
4 · Standardised measurements
Recurring, comparable tests: brightness/legibility under defined ambient light, verified weight, real battery run, assistant latency, pairing reliability.
5 · Separated scoring
We strictly separate measured value, cited third-party test and editorial assessment. Every grade states the methodology version and the test date.
6 · Correction & re-scoring
Software changes glasses. Major updates can lead to a re-scoring; we correct errors openly and mark the change.
Independence & funding
No paid grades. Manufacturers can buy neither a test nor a better rating.
No advance approval. Manufacturers do not see grades before publication and cannot change them.
Editorial team separated from sales. That we sell the Even G2 in Switzerland is disclosed — it gives the glasses no bonus in the score. They are measured according to the same protocol as any other pair and must earn their grade.
Disclosure. Sourcing, any affiliate links and conflicts of interest are named with every test.
Limits of this methodology
We are transparent, also about what the score does not deliver: it is an informed but subjective editorial assessment based on defined measurement points — not a laboratory certificate. Individual device units can vary. Software changes. And weightings are a deliberate value judgement that others might make differently. That is precisely why we disclose them — so that you can put our grade into context and, where you weight things differently, adjust it yourself.
Frequently asked questions
What is the AI Eyewear Score?
An independent overall grade from 0 to 100 for AI glasses, formed from six weighted criteria: Display & optics, Comfort & design, AI/software & updates, Battery & endurance, Privacy & transparency and Value for money. It is derived exclusively from tests we conduct ourselves.
Have any glasses been graded yet?
No. As of July 2026 only the methodology is published. We only assign a grade once we have tested a device ourselves according to this protocol — starting with the Even G2, which we continuously assess in our showroom.
Do manufacturers influence the grade?
No. Manufacturers do not see grades in advance, cannot change them and do not pay for tests or ratings. How a device was sourced is disclosed. The editorial team is separated from commercial operations.
Why do you sell the Even G2 and score it at the same time?
We disclose this conflict of interest. The Even G2 receives no bonus in the score; it is measured according to the same protocol and the same weighting as any other pair. Where it is weaker, we write that.