Advertised context length is a useful operational claim, but the current context-window surface is diagnostic only.
Algorithm
When enabled, send graduated long-context prompts with a recoverable needle and record whether the endpoint accepts the input and recovers the needle. Otherwise report the probe as not enabled.
Thresholds
Condition
Verdict contribution
Needle recovered at advertised window
Diagnostic match
Endpoint errors or loses the needle before the advertised window
Diagnostic anomaly
Any result
Score contribution remains 0
Limitations
Large-context probes are expensive and often disabled. Gateway cost caps, truncation policy, regional limits, or request-size limits can all explain failures without implying model substitution.
Anysingle signal cannot provemalicious behavior. Proxies may show anomalies for legitimate reasons (regional routing, A/B testing, degradation strategies, cache optimization).
Token ratio deviation may result from ChatML wrapping, system prompt injection, or tokenizer version differences — not necessarily intentional inflation.
Model identity judgment is based on statistical fingerprint matching, not cryptographic proof. Quantization, fine-tuning, and post-processing can all alter fingerprints.
MMD distribution tests are sensitive to temperature, sampling parameters, and system prompts. Significant p-values mean distributional difference, not proof of substitution.
Logprobs unavailability is increasingly common (many providers disable it by default in 2025-2026) and does not by itself indicate deception.
ITT rhythm fingerprinting is an early-stage technique. Network jitter, TCP coalescing, and gateway buffering can produce false signals.
This tool generates reference-grade evidence chains, not legal conclusions. Do not make definitive accusations based solely on this report.
The wording in the report refers to statistical "deviations" or "signal inconsistencies". Please do not use this to make fraud or deception claims against any service provider.