AI agent cost calculator
Estimate agent workflow costs across planner calls, executor calls, retries, fallback chains, and tool usage risk.
Last updated:
Scenario calculator
Calculate and compare
Planner/executor agent with retries and tool risk.
Gemini 3.5 Flash
Based on 12 months
Per active user / month
Per modeled task
$56.5 / month
$56.5 / month
Gemini 3.5 Flash is watchlist; verify sources before procurement. Cached input pricing is not confirmed for this model and is excluded from the estimate. Llama 3.1 70B Instruct is aggregator verified; verify sources before procurement.
| Model | Provider | Source status | Input / 1M | Output / 1M | Monthly cost | Cost / task |
|---|---|---|---|---|---|---|
| Gemini 3.5 Flash | Watchlist | $0.08 | $0.30 | $56.5 | $0.0023 | |
| Llama 3.1 70B Instruct | Meta / Hosted APIs | Aggregator verified | $0.59 | $0.79 | $261.63 | $0.0105 |
| GPT-4o | OpenAI | Official stale | $2.5 | $10 | $1,903.5 | $0.0761 |
| Claude 3.5 Sonnet | Anthropic | Official stale | $3 | $15 | $2,585.52 | $0.1034 |
Cache impact: 10% ยท selected models: 4
Hidden cost checklist
- Multi-step tasks
- Tool calls
- Retry loops
- Fallback models
Related planning pages
Compare this scenario against the main AI model cost calculator, provider pages, and the pricing change log before locking a production budget.
AI agent cost calculator FAQ
What assumptions does the AI agent cost calculator use?
The default preset starts with 150,000 monthly requests, 2,000 input tokens, and 700 output tokens per request.
Why does source status matter?
AI model prices change quickly. ModelMeter keeps watchlist, stale, aggregator, and official records visibly separate so estimates do not look more certain than the sources allow.