Refactor: Pricing/Draft-Pfad als Single Source of Truth (Phase 1)

- Neuer services/pricing.py: PRICING-Dict + compute_savings() aus dem
  system-Router extrahiert; Router ist jetzt dünn (nur role_map + Aufruf).
- /system/token-stats liefert zusätzlich das pricing-Dict → Frontend zeigt
  die Tarife daraus an statt sie im Text zu hartkodieren.
- SPEC_DRAFT_MODEL_PATH in config.py (MC_SPEC_DRAFT_MODEL); llamaswap.py und
  migrate_config.py referenzieren die Konstante statt des doppelten Literals.
- Ersparnis-Berechnung verhaltensneutral verifiziert (35,09 $ / 32,28 €).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

backend/routers/system.py

@@ -5,6 +5,7 @@ Lokal (Windows) schlagen die Shell-Befehle harmlos fehl und werden als Fehler
 zurückgegeben statt zu crashen.
 """
 
+import logging
 import os
 import subprocess
 
@@ -15,8 +16,12 @@ from config import GATEWAY_URL, HERMES_API_URL, HERMES_WEBUI_URL, LLAMA_SWAP_URL
 from services import backup as backup_svc
 from services.agent import agent_status
 from services.gateway import gateway_reachable
-from services.llamaswap import engine_reachable
+from services.llamaswap import engine_reachable, list_models
+from services.pricing import compute_savings
 from services.system import system_status
+from services.token_stats import get_stats
+
+log = logging.getLogger(__name__)
 
 router = APIRouter(prefix="/api")
 
@@ -90,64 +95,16 @@ def self_update() -> dict:
     return {"pull": pull, "reset": reset, "restart": restart_res}
 
 
-from services.token_stats import get_stats
-from services.llamaswap import list_models
-
 @router.get("/system/token-stats")
 def token_stats() -> dict:
-    stats = get_stats()
-    p = stats.get("prompt_tokens", 0)
-    c = stats.get("completion_tokens", 0)
-    total = p + c
-    
-    # Map model IDs and aliases to their respective roles for pricing resolution
-    role_map = {}
+    """Token-Verbrauch + Cloud-Ersparnis. Logik im pricing-Service (SSoT)."""
+    # Rolle je Modell/Alias (lowercase) für die Tarif-Auflösung auflösen.
+    role_map: dict[str, str | None] = {}
     try:
         for m in list_models():
             role_map[m["name"].lower()] = m.get("role")
             for alias in m.get("aliases", []):
                 role_map[alias.lower()] = m.get("role")
     except Exception:
-        pass
-
-    # Dynamic pricing tiers based on model class in June 2026
-    PRICING = {
-        "heavy": (15.0, 75.0),
-        "coder": (3.0, 15.0),
-        "hermes": (1.0, 5.0),
-        "fast": (0.15, 0.60),
-        "scout": (0.15, 0.60),
-        "vision": (0.15, 0.60),
-        "reasoning": (0.15, 0.60),
-    }
-
-    modeled_p = 0
-    modeled_c = 0
-    saved_usd = 0.0
-    
-    models_data = stats.get("models") or {}
-    for m_name, m_tokens in models_data.items():
-        mp = m_tokens.get("prompt", 0)
-        mc = m_tokens.get("completion", 0)
-        modeled_p += mp
-        modeled_c += mc
-        
-        role = role_map.get(m_name, m_name)
-        rate_in, rate_out = PRICING.get(role, (0.15, 0.60))
-        saved_usd += (mp * rate_in + mc * rate_out) / 1_000_000.0
-
-    # Baseline/legacy tokens calculated at premium rates ($15.00 / $75.00)
-    # to preserve historical savings value prior to model-specific logging
-    baseline_p = max(0, p - modeled_p)
-    baseline_c = max(0, c - modeled_c)
-    saved_usd += (baseline_p * 15.0 + baseline_c * 75.0) / 1_000_000.0
-    
-    saved_eur = saved_usd * 0.92  # 1 USD = 0.92 EUR
-    
-    return {
-        "prompt_tokens": p,
-        "completion_tokens": c,
-        "total_tokens": total,
-        "saved_usd": round(saved_usd, 2),
-        "saved_eur": round(saved_eur, 2)
-    }
+        log.warning("token_stats: list_models fehlgeschlagen, Tarife per Name", exc_info=True)
+    return compute_savings(get_stats(), role_map)