Directory evidence
Factual listing review
Free
For a model/API page, provider profile or correction supported by official documentation.
- Provider and model identity
- Current pricing source
- Access and lifecycle notes
- Explicit unavailable fields
Home / Submit an API for review
For LLM API providers
Submit the evidence. Not a preferred conclusion.
A factual listing review is free. A custom benchmark study is considered only after the workload, protocol, dataset, disclosure and publication rules are written down.
Direct answer: Payment cannot buy a better score or ranking. If a measured result is commissioned, the method and result remain independent—even when they are unfavorable.
Two paths
Correct the directory or scope original research.
No fixed benchmark price is published before the measurement plan exists. That would sell an outcome before defining the evidence.
Original research
Custom benchmark study
Scoped
For a defined workload that can be tested reproducibly and published with its limitations.
- Preregistered protocol
- Versioned prompts and environment
- Request-level provenance
- Independent result and disclosure
Publication rules
The evidence controls the page.
These rules apply whether or not commercial work follows.
01
No pay-to-rank
Commercial status cannot change factual tables or conclusions.
02
Method first
Metrics and pass criteria are defined before measurement begins.
03
Gaps stay visible
Missing or failed coverage is not converted to a favorable zero.
04
Disclosure in place
Commissioned research is labelled where readers see the result.