Overhaul model system with multi-stage training, variants, benchmarks, and eval

Replace the single-stage training + flat capability score with a realistic AI
development pipeline: pre-training with Chinchilla scaling laws, SFT with
specializations, alignment with safety/capability tradeoffs (RLHF/DPO/Constitutional),
model families with distillation/fine-tuning/quantization variants, named benchmark
suite with compute-costing eval jobs, and segment-specific market quality.

Phases 1-6 of the model rework plan: new types, engine rewrite, save migration,
training events/risk system, concurrent training, variant creation, benchmark
evaluation with leaderboard, and market integration.

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

packages/game-engine/src/systems/eraSystem.ts

@@ -11,9 +11,7 @@ export function checkEraTransition(state: GameState): Era | null {
   const thresholds = ERA_THRESHOLDS[nextEra as keyof typeof ERA_THRESHOLDS];
   if (!thresholds) return null;
 
-  const bestModel = state.models.trainedModels.reduce(
-    (best, m) => Math.max(best, m.benchmarkScore), 0,
-  );
+  const bestModel = state.models.bestDeployedModelScore;
 
   if (
     state.economy.totalRevenue >= thresholds.revenue &&