token-budget-trackerlisted

# Token Budget Tracker ## Overview Monitor AI model usage, costs, and optimize agent budget allocation. Provides visibility into real spend per agent, per task, and per day. ## Why Track Tokens | Problem | Impact | Solution | |---------|--------|----------| | Blind spend | Budget overrun | Real-time cost visibility | | Inefficient routing | Paying GPT-4 for simple tasks | Model router optimized per task | | No accountability | Can't attribute costs to outcomes | Per-agent cost tracking | ## Current Model Costs (as of Feb 2024) | Model | Provider | Input | Output | Use Case | |-------|----------|-------|--------|----------| | **Kimi K2.5** | OpenRouter | $3/M tokens | $3/M tokens | Most tasks, default | | **Claude Sonnet 4.5** | Anthropic | $15/M tokens | $75/M tokens | Complex reasoning, content | | **Claude Opus 4.5** | Anthropic | $75/M tokens | $150/M tokens | Deep research only | | **GPT-4 Turbo** | OpenAI | $30/M tokens | $60/M tokens | Fallback, legacy | | **Grok 3** | xAI | $10/M tokens | $20/M tokens | X/Twitter research | ## Budget Allocation (Monthly) ```yaml total_budget: $500/month allocation: axioms_main: $150 # 30% - Coordinator, main interface trader: $100 # 20% - Trading decisions (fast, frequent) creator: $120 # 24% - Content generation (creative) investment: $80 # 16% - Research (thorough, deep) mission_control: $50 # 10% - Overhead, tracking cost_optimization: # Use cheaper models for routine tasks heartb