Agent IRI that provides this model.
Current budget ratio (remainingTokens / maxTokens), if a budget was
queried. null if no budget was provided or the graph returned no row.
mesh:ModelCapability node IRI.
Maximum token context window.
USD cost per output token.
Whether the downgrade path was applied.
true means the cheapest model was selected because the remaining token
budget fell below the configured threshold.
Provider-specific model identifier (e.g. 'gpt-4o').
Organisation providing the model (e.g. 'openai').
Normalised quality score in [0.0, 1.0].
A resolved model capability, returned by selectModel.