Observed community preference and recorded token usage for each model and effort setting
·
Preference = (wins + ½ ties) ÷ ratings. “Cannot judge” is excluded. Early scores depend on which opponents each model faced. This measures visual preference, not BIM correctness. † Includes reported helper agents; other agent scopes are unspecified. Token counts are not billed cost.
Exact token counts & sources+
Token counts for the selected configurations
Model · Effort
Total
Input
Cached input¹
Output
Agent scope
PDF
¹ Cached input is part of input; reasoning is part of output. — = unreported. * = arithmetic derivation. GPT-6 Sol · Ultra uses the reported combined total and combined cached count; input/output are unreported. Its separate chat-only snapshot is 7,119,141.
One result per model and effort setting. The PDFs show the visible drawing; native element identity, editability and execution-rule compliance have not been verified. Token figures are snapshots, with differing agent scopes. Community preferences describe these drawings; no native BIM score, billed cost or run time is inferred.