The author manages a daily workflow of over 1 billion tokens using various AI subscriptions. To optimize costs, she employs a strategy of delegating routine tasks to smaller, cost-effective models like GPT 5.6 Luna, Qwen 3.8 Flash, and Deep Seek V4 Flash, while reserving more powerful models for complex requirements.
The GPT Pro subscription, priced at $100, provides 7–9 daily streams of Luna and 50 weekly requests to GPT Astra Pro. Users can integrate these with Google Drive to manage complex tasks and maintain organized project folders, effectively bypassing some limitations of individual web interfaces.
Minimax offers high speed and large limits within a 5-hour window, making it suitable for routine tasks. Meanwhile, Kimi subscriptions start at $40, which the author considers less efficient compared to alternatives like Qwen or Ollama Cloud. Ollama Cloud provides a transparent model where a $20 payment grants $60 in token credits.
Grok is highlighted for its video generation capabilities, offering 400 clips of 15 seconds in 720p resolution for $30. For budget-conscious users, OpenCode Go serves as an emergency subscription for $10. Additionally, the author uses Krea for image generation at $9 per month and pays $5 for Minimax audio packages to improve voice synthesis quality.
The total monthly expenditure for the author's current setup, including GPT Pro, Minimax, and OpenCode Go, amounts to $130. This configuration allows for handling high-volume AI tasks while balancing performance and cost across different specialized platforms.
