Caller model routing — cycle, style-matched, mid-show override

- Three strategies: single model, cycle through pool, style-matched
- 18 communication styles mapped to 7 models (Grok, Sonnet, Mistral, Qwen, DeepSeek, Gemini, Llama)
- Per-caller model locked for entire call, overridable mid-show
- Model badges on caller buttons and info panel
- Settings UI for strategy, pool, style mapping, fallback
- Fallback to Sonnet on model failure
- 6 new models added to pricing and dropdown
- Checkpoint persistence for all model state

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

backend/services/cost_tracker.py

@@ -45,6 +45,12 @@ OPENROUTER_PRICING = {
     "openai/gpt-4o-mini":              {"prompt": 0.15,  "completion": 0.60},
     "openai/gpt-4o":                   {"prompt": 2.50,  "completion": 10.00},
     "meta-llama/llama-3.1-8b-instruct": {"prompt": 0.06, "completion": 0.06},
+    "deepseek/deepseek-chat-v3-0324":  {"prompt": 0.27,  "completion": 1.10},
+    "moonshotai/kimi-k2":              {"prompt": 0.60,  "completion": 2.00},
+    "mistralai/mistral-medium-3":      {"prompt": 0.40,  "completion": 2.00},
+    "meta-llama/llama-4-maverick":     {"prompt": 0.20,  "completion": 0.60},
+    "qwen/qwen3-235b-a22b":           {"prompt": 0.20,  "completion": 0.60},
+    "google/gemini-2.5-pro":           {"prompt": 1.25,  "completion": 10.00},
 }
 
 # TTS pricing per character