# Plan Quotas, Unified $10/50k Auto-Scale & LLM Token Savings Math

> undefined

## 1. Plan Quotas & Unified Flat Auto-Scale Overage ($10 per 50,000 Checks)

- **Starter ($0 / Free Trial):** 250 included checks ($0.00). No Auto-Scale overages allowed; silently fails open at 250 checks so your app never breaks.
- **Builder ($19 / month):** 25,000 included checks/month ($0.00076 / check). Supports optional, customer-controlled Auto-Scale overage at a flat **$10 per 50,000 additional checks ($0.00020 / check)**.
- **Business ($49 / month):** 250,000 included checks/month ($0.000196 / check), exclusive access to `GET /v1/analytics/export` (5-minute polling interval), e-commerce coupon/refund exploit blocking, and the same flat **$10 per 50,000 additional checks ($0.00020 / check)** Auto-Scale rate.

## 2. Cost Per Check vs. Downstream LLM & AI Builder Token Costs

| Downstream Model / Builder | 1 Normal Moocher Turn (~2k in / 800 out) | 1 Paste-Bomb Attack (~30k in / 2k out) | Gated Checks Paid For by 1 Blocked Bomb |
| --- | --- | --- | --- |
| **AI Builder Credits (Lovable / Replit / Bolt / v0 / Vercel)** | $0.0500 (500x overage check cost) | **$0.2500** | **500 to 2,500 checks** |
| **Claude Sonnet (3.7 / 4)** ($3/1M in, $15/1M out) | $0.0180 (180x overage check cost) | **$0.1200** | **1,200 checks** |
| **OpenAI GPT-4o** ($2.50/1M in, $10/1M out) | $0.0130 (130x overage check cost) | **$0.0950** | **950 checks** |
| **Google Gemini 2.5 Pro & Mistral Large** | $0.0110 (110x overage check cost) | **$0.0680** | **680 checks** |
| **Reasoning Models (OpenAI o3 / Claude Opus / DeepSeek R1)** | $0.0780 (780x overage check cost) | **$0.5700** | **5,700 checks** |
| **Open-Weight Frontier (Llama 3.3 70B / DeepSeek V3 / Qwen 2.5 via Together, Fireworks, Groq, OpenRouter)** | $0.0030 (30x overage check cost) | **$0.0240** | **240 checks** |
| **Cloud Managed Open-Weights (AWS Bedrock / Azure / Cloudflare Workers AI)** | $0.0050 (50x overage check cost) | **$0.0380** | **380 checks** |

## 3. Breakeven Abuse Thresholds

- **Builder ($19/mo, 25,000 checks):** Blocking just **160 paste-bombs** all month (**0.64%** of checks) saves $19.20 on Claude Sonnet and pays for the entire monthly plan.
- **Business ($49/mo, 250,000 checks):** Blocking just **410 paste-bombs** (**0.16%** of checks) or stopping **1 single $50 coupon exploit** pays for the entire month.
- **Auto-Scale Pack ($10 per 50,000 checks):** Stopping just **84 paste-bombs** out of 50,000 checks (**0.168%** of traffic) saves $10.08 and pays for the entire 50k pack. At a typical 3% bot/abuse rate (1,500 blocked turns), a $10 pack saves **$27 to $180** in direct LLM spend.