YFarmXIntelligence for an accelerating worldBriefings
YFarmX
YFarmX
AI Desk · Model Identity

Know what is behind the endpoint

An independent evidence layer for AI models. The label says what the provider calls a model; YFarmX measures the tokenizer underneath, checks the serving contract, and keeps every observation as dated history.

Try:

Models indexed
486
Measured
280
Complete fingerprints
262
Tokenizer groups
51
Unlabelled models placed
76 of 125
New since 3 Sept
34

Catalogue and measurements as at 18 September 2026. Measured counts every model with a validated measurement; complete fingerprints are the subset whose 50 rows all came back clean, and only those carry an exact signature. Every figure derives from one committed dataset; the methodology records how it is measured, and what the method cannot see.

Live discovery

A stealth model, measured while its identity stays undeclared

What the measurements say

Findings the current dataset supports
Verified findingDeepSeek: R1 Distill Llama 70B: the bill follows a vocabulary the model does not readDeepSeek's Llama distill reads with Llama 3's tokenizer, verified from its own repository files, while the endpoint's reported token counts, the numbers customers pay on, follow DeepSeek's vocabulary. The verification note on its page carries the evidence.See the evidence →

Explore

Every modelAll 486 endpoints on one page. Open any row for its evidence: what the measurement found, how sure we are, and what shares its fingerprint.Open the catalogue →CompareTwo models, side by side: tokenizer agreement string by string, API surface field by field.Open the comparison bench →Identity watchNew catalogue arrivals, fingerprint changes and fresh lineage placements, as they are observed.See what changed →UnknownsStealth models whose maker stays hidden, and the 125 models with no family on their label. Measurement fills the blanks: 76 placed so far.Open the queue →Tokenizer AtlasThe tokenizer-family structure of the catalogue: clusters, constellation and the family explorer.Open the Atlas →MethodologyHow the fingerprints are measured, the controls that gate publication, and the known limits.Read the method →

How it works

01 · MEASURE

Fifty adversarial strings are sent to each model and only the reported prompt-token count is read. The marginal cost of each string under the model's tokenizer becomes a 50-number signature.

02 · COMPARE

Signatures are compared offline against every other model measured. Exact agreement across all fifty strings is strong lineage evidence; the API-surface contract is checked as a second signal of a different kind.

03 · TRACK

Every observation is kept as taken, and the history only ever appends. When an endpoint's fingerprint moves, the history shows what changed and on which day.

A shared tokenizer is evidence of shared lineage, never proof of an exact model. Theclaims hierarchy keeps every published conclusion inside what the evidence supports, and the controls are published with the results.