It is distinct from the full internal operating methodology and does not disclose proprietary weights, calibration parameters or internal implementation procedures.
This statement is prepared for public disclosure of the methodology. Its sections explain evaluation principles but do not provide a complete specification for reproducing the calculations.
DISCLOSURE SCOPE The evaluation subjects, dimensions, factual and evidentiary standards, and limitations are explained. Proprietary weighting, internal calculation procedures, calibration models and other implementation materials are not disclosed.
1. Rating Purpose and Scope
QiaoLink Rating independently evaluates products, services and other works developed with AI assistance or incorporating AI during operation. Its ratings provide a reference for adoption, procurement and replacement decisions by users, developers and relevant institutions.
The evaluation focuses on actual value in a defined use case, rather than technological sophistication, model size, market prominence or fundraising activity.
QiaoLink Rating covers two categories of evaluated products and works:
Category A | AI-in-Operation Products and Services. Products and services in which artificial intelligence participates in judgment, generation, processing or task execution during operation, including AI agents, intelligent automation systems and related applications.
Category B | AI-Assisted Products and Works. Products, tools, applications, research works and solutions developed, designed or produced with AI assistance, where the completed work does not necessarily rely on AI during operation.
Categories A and B identify the type of evaluated product or work; they do not indicate a higher or lower rating.
QiaoLink Rating forms its conclusions with reference to the intended users, principal use case, practical adoption conditions and available alternatives. The scale, industry or commercial maturity of an evaluated product or work does not, by itself, determine its final rating.
2. Ten-Dimension Evaluation Framework
QiaoLink Rating applies a uniform ten-dimension evaluation framework. Each dimension has a defined scope and contributes to the assessment of practical adoption value.
| Dimension | Evaluation Area | Basic Evaluation Principle |
|---|---|---|
| D1 | Task Completion Quality | Assesses the correctness, completeness and actual deliverable quality with which the evaluated product or work performs its intended task. |
| D2 | Efficiency / Productivity Improvement | Assesses the actual effects of adopting the evaluated product or work on time, operating steps, human effort and output efficiency. |
| D3 | Stability | Assesses reliability under repeated use, extended operation and changing operating conditions. |
| D4 | Human Effort After Adoption | Assesses human operation, supervision, review and correction required during normal use. |
| D5 | Deployment / Adoption Barriers | Assesses the conditions target users must meet to obtain, integrate, learn and adopt the evaluated product or work in practice. |
| D6 | Adoption Cost and Economics | Assesses monetary costs over the practical adoption lifecycle and their economic reasonableness relative to the value obtained. |
| D7 | Maintenance Cost | Assesses the technical fixes, updates and adaptations required to keep the evaluated product or work usable. |
| D8 | Reuse / Portability / Scalability Value | Assesses the practical value of reusing, transferring to new environments and further extending the evaluated product or work and its creation methods. |
| D9 | Limitations / Risk / Controllability / Reversibility | Assesses risks arising from adoption, practical control capabilities and the ability to correct, take over or exit safely when necessary. |
| D10 | Incremental Value over Existing Alternatives | Assesses the actual additional value offered relative to alternatives realistically available to target users. |
The ten dimensions apply to both Category A and Category B products and works. The specific evidence used must reflect how the evaluated product or work operates, its use case and the evidence available for verification.
Where an evaluation item is genuinely inapplicable, it is handled under the applicability provisions of the formal methodology. Inapplicability is not treated directly as a defect.
3. Non-Equal Weighting Principles
QiaoLink Rating employs a ten-dimension, non-equally weighted evaluation mechanism. The dimensions do not have identical effects on the final score, and QL Score is not a simple arithmetic mean of the ten dimension scores.
The base weight of each dimension is determined in advance by the unified methodology. The same methodology version applies consistent base-weighting rules. These rules are not changed ad hoc on account of the developer’s identity, commercial relationships, market prominence or a desired rating outcome.
Each dimension is assessed against actual evidence and applicable conditions. Cases involving insufficient evidence, inapplicable evaluation items or particular risks are handled under established rules.
Specific dimension weights and internal scoring and calibration parameters are proprietary QiaoLink methodological materials and are not included in this public document.
Accordingly, this statement explains the basis and principles of formal ratings; it is not a complete computational specification from which QL Score can be independently reproduced.
4. Evidence Classification, Verification and Anti-Manipulation
QiaoLink Rating forms its judgments from verifiable information and distinguishes among the source, strength and scope of verification of evidence.
4.1 Evidence Classification
| Evidence Class | Basic Meaning |
|---|---|
| E0 | Insufficient information; no adequate verifiable basis is currently available. |
| E1 | Statements or materials provided by creators, developers or project operators. |
| E2 | Facts that QiaoLink can directly observe, verify or confirm by returning to the original source. |
| E3 | Reliable, independent and traceable external corroboration. |
Evidence classes indicate the basis on which information is verified. They do not, in themselves, represent a quality grade for the evaluated product or work.
4.2 Evidence Verification and Factual Judgment
Public statements, performance claims or commercial figures from a project do not automatically constitute independent evidence merely because they have been published.
Repeated citations of the same original information by different sources do not necessarily constitute multiple independent pieces of evidence.
Insufficient evidence must be distinguished from a verified defect.
A lack of adequate verification does not, by itself, establish a quality problem. Equally, the absence of evidence supporting a critical capability cannot justify treating that capability as proven to a high standard.
Where material evidence gaps affect an adoption decision, the rating report should explain their nature and implications for the conclusion.
Unverified rumors, speculation or negative allegations are not treated as established facts and may not, on their own, constitute a basis for lowering a rating.
4.3 Evidence Anti-Manipulation Mechanism
For reviews, ratings, likes and other external feedback potentially subject to human manipulation, QiaoLink maintains internal evidence-verification and anti-manipulation mechanisms.
Where the authenticity or independence of anomalous information has not been sufficiently verified, QiaoLink does not treat such information as a decisive basis for a rating, nor does it publicly conclude on that basis that the parties concerned have engaged in manipulation.
5. QL Score and QL Rating
QiaoLink Rating presents its conclusions through QL Score and QL Rating.
5.1 QL Score | Composite Score
QL Score is expressed as an interval on a 0–100 scale, representing the composite evaluation of a product or work under its defined use case and the available evidence.
The interval comprises a lower and upper bound and represents the reasonable range supported by the evidence currently available.
It is not a probability distribution of actual product performance, nor does it imply that all values within the interval are equally likely.
Scores are determined under the formal methodology’s uniform scoring and calibration rules. Publicly displayed values may be formatted for presentation; formal rating grades are determined using unrounded computational results and applicable rating rules.
5.2 QL Rating | Composite Rating
QL Rating is determined from the composite score and the applicable conditions specified in the formal methodology. The rating scale is:
| Rating | Meaning |
|---|---|
| A+ | Exceptional — Highest Recommendation |
| A | Highly Recommended |
| A- | Recommended for Adoption |
| B+ | Good |
| B | Meets the Standard |
| B- | Meets the Basic Standard |
| C+ | Adopt Under Limited Conditions |
| C | Minimum Threshold for Considering Adoption |
| Does Not Meet the C-Level Standard | Does Not Meet the Minimum Adoption-Value Standard |
Each grade reflects the composite adoption value of the evaluated product or work under specific conditions. It is not a direct assessment of the developer’s ability, enterprise value or future commercial performance.
Rating grades are determined using uniform numeric mapping and any necessary special application rules. Specific numerical thresholds for grade assignment and internal calibration parameters are not disclosed in this public statement.
5.3 NR | Not Rated
When reliable evidence currently available is insufficient to support a credible composite-score interval, QiaoLink may classify the result as NR (Not Rated).
NR is a separate evaluation status, not a lower rating grade, and does not indicate that the evaluated product or work has been proven to have material defects.
For products or works assigned NR, QiaoLink will identify in the public report the principal evidence gaps preventing a rating. Where relevant critical evidence is materially supplemented, QiaoLink may undertake a new evaluation under the formal methodology.
6. Risk Evaluation and Rating Limitations
Risk evaluation is grounded in the actual use case of the evaluated product or work and considers the consequences of risk, conditions of occurrence, detectability, ability to correct, available controls and exit conditions.
Industry classification alone does not determine risk severity.
The same product or work may have different risk characteristics depending on permissions, operating environment or human-supervision conditions.
Where reliable evidence establishes a major risk, QiaoLink applies the risk-treatment provisions of the formal methodology to limit the resulting rating. Where further verification confirms an unacceptable risk, the appropriate determination is made under the formal rating rules.
Risk treatment distinguishes established facts from reasonable inferences and unverified information.
Risk conclusions must correspond to the applicable use case and scope of evidence. Major risk must not be inferred solely from the industry in which a product operates or from unverified negative information.
7. Rating Outlook
Rating Outlook reflects the possible direction of changes over the next three months in conditions relevant to the current rating.
Rating Outlook communicates information distinct from the current QL Rating. An outlook does not automatically trigger a change in the formal rating.
QiaoLink uses the following outlook designations:
| Outlook | Meaning |
|---|---|
| Positive | Available evidence supports the possibility of favorable changes in relevant conditions. |
| Stable | Available evidence does not indicate a trend sufficient to support a clear directional change. |
| Under Pressure | Relevant developments may put pressure on the future adoption value of the evaluated product or work. |
| Developing | Important changes have occurred, but their combined direction of impact remains uncertain. |
| No Outlook | Comparable data or reliable evidence is insufficient to support an appropriate directional assessment. |
Outlook assessments consider materially significant updates to the evaluated product or work, external participation and other verifiable developments.
Trend comparisons should use comparable time periods and consistent data definitions. Short-term promotional campaigns or unverified changes in attention are not, by themselves, treated as sustained development trends.
Developing is not synonymous with insufficient evidence; No Outlook does not imply a negative assessment.
8. Factors That Could Lead to an Upgrade
Formal rating reports include a “Factors That Could Lead to an Upgrade” section, stating specific conditions that could support a higher rating in the future. Where this section is not applicable, that fact is noted.
Upgrade factors must be directly relevant to existing limitations, material evidence gaps or the performance of the evaluated product or work and must be supported by verifiable grounds.
New features, user growth or supplementary materials submitted by a project operator do not automatically justify an upgrade.
A new rating conclusion may be reached only after the relevant developments have undergone necessary verification and a new evaluation has been completed under the formal rating rules applicable at that time.
Stated upgrade factors do not constitute a commitment to change a rating in the future.
9. Rating Independence and Review Principles
QiaoLink Rating evaluates the actual adoption value of a product or work on the basis of verifiable evidence.
Promotional reach, creator prominence, market attention or commercial relationships do not substitute for factual assessment.
An evaluated party cannot alter formal rating rules by specifying a desired grade in advance.
Formal rating conclusions are published after evidence verification and a consistency review of the rating. Material factual errors, significant changes in available information or substantive changes in the basis of assessment are handled through the applicable correction or re-evaluation procedures.
Formal rating reports set out the principal factual basis, key reasoning and applicable limitations.
Where facts are disputed, QiaoLink distinguishes verified information from matters still requiring verification. A differing view from a source of information or another third party does not automatically overturn an existing rating.
Rating changes must be grounded in verifiable changes in facts, updated evidence or a formal methodological change.
10. Methodology Version and Scope of Application
This public statement corresponds to QiaoLink Rating Unified Formal Methodology v3.2, dated 22 October 2026.
QiaoLink identifies the rules underlying a rating by methodology version. Methodological changes are recorded with the relevant version and effective date.
Rating reports issued on different dates should be interpreted in light of the methodology version applicable at their issuance, the state of the evaluated product or work and the available evidence.
QiaoLink retains its internal scoring model, weighting parameters, calibration rules and implementation materials. This public statement discloses the principal evaluation principles but does not replace the full internal operating methodology.
QL Rating, QL Score and Rating Outlook are independent evaluative judgments made for a particular time, use case and set of evidentiary conditions.
Ratings do not guarantee future performance, investment returns, legal compliance or specific outcomes of use. Adoption decisions involving medical, financial, legal or other specialist matters must also take account of the relevant professional requirements.