fix: raise heavy threshold — 4000→12000 tokens, 8→15 turns

Agent conversations with system prompts easily exceed 4000 tokens,
forcing everything to Dense. Now only truly heavy work triggers Dense.
Most agent convos will route to MoE (default) instead.

router/router.py

@@ -181,8 +181,8 @@ def route(rd, tier):
         if "qwen3.6-27B-code" in avail and not is_gpu_busy("qwen3.6-27B-code"):
             return {"model":"qwen3.6-27B-code","reason":"simple_conv_fallback"}
     
-    # TIER 3: Heavy reasoning — large context or very long conversations
-    if t > 4000 or turns > 8:
+    # TIER 3: Heavy reasoning — very large context or very long conversations
+    if t > 12000 or turns > 15:
         candidates = [m for m in ["qwen3.6-27B-code","qwen3.6-35B-A3B","qwen3.5-9b-vlm"] if m in avail]
         result = select_best_gpu(candidates, "heavy_reasoning")
         if result: return result