Added the Model Comparison assistant for blind comparisons of two models (#1009)

Co-authored-by: Thorsten Sommer <SommerEngineering@users.noreply.github.com>

app/MindWork AI Studio/Assistants/ModelComparison/ModelComparisonModelSpeed.cs

@@ -0,0 +1,32 @@
+namespace AIStudio.Assistants.ModelComparison;
+
+/// <summary>
+/// How fast a model answered across all stored comparisons.
+/// </summary>
+/// <remarks>
+/// Medians rather than means: a single answer which waited a minute for a busy server would otherwise shift the whole
+/// picture. The times depend on the length of the answers and the load of the server as much as on the model, so they
+/// compare models only when both ran on the same tasks.
+/// </remarks>
+public sealed record ModelComparisonModelSpeed
+{
+    /// <summary>
+    /// The model, with the name it had in its latest comparison.
+    /// </summary>
+    public required ModelComparisonModelIdentity Model { get; init; }
+
+    /// <summary>
+    /// How many complete answers the times come from.
+    /// </summary>
+    public int AnswerCount { get; init; }
+
+    /// <summary>
+    /// The median time until the first piece of an answer arrived, or null without any.
+    /// </summary>
+    public TimeSpan? MedianFirstTokenTime { get; init; }
+
+    /// <summary>
+    /// The median time until an answer was complete, or null without any.
+    /// </summary>
+    public TimeSpan? MedianTotalTime { get; init; }
+}