← All patterns

Single LLM Call

One prompt in, one response out. No tools, no loops.

API Calls per Conversation

Low

1

Mid

1

High

1

Best Use Cases

Simple Q&A chatbotsText classificationContent generation (one-shot)Summarization

Cost Drivers

  • Token count is the only variable — no tool overhead, no loops
  • Cost scales linearly with conversation volume
  • Model choice is the biggest lever (15x range between Haiku and Opus)

Monthly Cost Scenarios

100/day

Haiku 4.5

$30-60/mo

100/day

Sonnet 4.5

$90-180/mo

1,000/day

Haiku 4.5

$300-600/mo

Optimization Tips

  • Use prompt caching if your system prompt exceeds 1,024 tokens
  • Route simple queries to cheaper models (Haiku or GPT-4o Mini)
  • Use batch API for non-real-time workloads (50% discount)

Estimate your single llm call costs

Describe your system, get a cost breakdown in 60 seconds. Free, no signup required.

Estimate Your System →