Home › Solutions › AI Agent: Cost per Task
Use caseAI Agent: Cost per Task
Agents resend their context on every step, so cost per task grows with steps. Caching the shared prefix and capping steps are the biggest levers.
Defaults use GPT-6.1 Sol's list prices ($2 input, $0.10 cached, $10 output per 1M tokens, checked October 30, 2026). Enter your own model's prices.
Why agent tasks are expensive
Each step resends the system prompt, tools and history. A 15-step task can pay for the same prefix 15 times unless it's cached.
How to cut it
- Cache the system prompt and tools.
- Cap steps per task.
- Trim tool results before they enter the context.
- Use cheaper models for simple steps.
Read how to cap AI agent costs.
Frequently asked questions
How much does an AI agent cost per task?
Steps per task times tokens per step times your model's prices. Caching the shared prefix and capping steps cut it the most.
Why are AI agents so expensive?
They resend their whole context on every step, so cost grows with the number of steps.
Want us to find your 20%?
Book a free 15-minute API cost teardown. Bring last month's provider usage and we'll show you your biggest waste and the first fix we'd ship.
Get a Free 15-Minute API Cost Teardown