Remove benchmark evaluation system, use training capabilities directly

Model quality for market segments and product lines now derives from deployed
model capabilities (coding, reasoning, agents, etc.) instead of requiring a
separate manual benchmark evaluation step. This eliminates an unbounded
benchmarkResults[] array that was scanned 5x per tick and removes ~480 lines
of dead-weight UI, types, and engine code.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

packages/game-engine/src/__test-utils__/builders.ts

@@ -171,7 +171,6 @@ export function createTestBaseModel(overrides?: Partial<BaseModel>): BaseModel {
     sizeTier: 'small',
     isPointRelease: false,
     sourceModelId: null,
-    benchmarkResults: {},
     dataMix: { web: 0.4, code: 0.2, books: 0.15, academic: 0.1, conversational: 0.1, specialized: 0.05 },
   };
   return overrides ? { ...base, ...overrides } : base;
@@ -181,9 +180,10 @@ export function createTestModelFamily(overrides?: Partial<ModelFamily>): ModelFa
   const base: ModelFamily = {
     id: uuid(),
     name: 'Test Family',
-    baseModels: [],
+    generation: 1,
+    baseModelIds: [],
     variants: [],
-    activeEvals: [],
+    createdAtTick: 0,
   };
   return overrides ? { ...base, ...overrides } : base;
 }