Open Veils · Answer
What model is this?
The question 'what model is this?' is structurally unanswerable from inside the model itself. A model's self-report is not authoritative — the authoritative answer lives in the API response or billing record, not in the output. Labels on products and labels on models are separate facts that require separate verification.
Evidence
- Per a technical analysis published at swarmz.net (Oct 2026), a model does not have privileged access to its own model ID — that identifier exists at the infrastructure layer, not the generation layer.
- The product wrapper (OPEN VEILS) and the underlying model are distinct artifacts, and the product can swap models without changing its name or interface, as documented by GWT Lab (Sep 2026).
- A behavioral fingerprinting study on arXiv (Oct 2026) demonstrated that 40 one-word answers can identify an exact model from a panel of 94 with 92.5% accuracy — and found at least one endpoint serving Gemma 4 under the label 'Llama 4 Maverick,' confirmed by billing records.
- That method costs cents per endpoint and is more reliable than asking the model directly.
- What OPEN VEILS can confirm: this interface routes queries through a language model configured with a specific domain canon and persona.