June 21, 2026
Here's What $1 of OpenRouter Credit Actually Gets You
Concrete examples of how far your money goes with pay-as-you-go AI through ZDR Chat.
One of the best things about ZDR Chat is that there’s no monthly subscription. You put money into your OpenRouter account, and it lasts until you spend it. But what does that actually look like in practice?
The Short Answer
$1 can cover weeks of regular use for most people. Here’s the breakdown.
Per-Million Token Pricing
OpenRouter charges per million tokens (roughly 750,000 words). Most models fall into these ranges:
| Model Tier | Input (per M tokens) | Output (per M tokens) | Example |
|---|---|---|---|
| Free | $0 | $0 | DeepSeek V3, Llama 3 |
| Budget | $0.15–$0.50 | $0.60–$1.50 | GPT-4o Mini, Claude Haiku |
| Mid-range | $1–$3 | $4–$15 | Claude Sonnet, GPT-4o |
| Premium | $3–$15 | $15–$75 | Claude Opus, GPT-5 |
A “token” is roughly ¾ of a word in English. A typical short message is ~50–100 tokens. A long message with context might be 500–2,000 tokens.
Real-World Examples
Example 1: Everyday Q&A ($0.10)
- Model: Claude Sonnet (mid-range)
- You ask: “Explain quantum computing in simple terms” — 20 tokens
- It replies: ~300 tokens
- Cost: $1.29/M input × 20 + $5.60/M output × 300 = $0.0017
That’s less than a fifth of a cent. For one dollar, you could ask and get answers to this question ~580 times.
Example 2: Drafting an Email ($0.005)
- Model: GPT-4o Mini (budget)
- You write: “Draft a professional email declining a meeting request politely” — 15 tokens
- It replies: ~150 tokens
- Cost: $0.15/M × 15 + $0.60/M × 150 = $0.00009
Practically free. You’d need to send ~11,000 emails to spend $1.
Example 3: Code Review (Moderate Usage)
- Model: Claude Sonnet
- You paste: 100 lines of code (400 tokens) + “Review this for bugs” (10 tokens)
- It replies: ~500 tokens of analysis
- Cost: $1.29/M × 410 + $5.60/M × 500 = $0.0033
About a third of a cent. $1 buys ~300 code reviews.
Example 4: Analyzing a 50-Page Document (Heavy Usage)
- Model: Claude Opus (premium)
- You upload: 50 pages of text (~50,000 tokens for context + ~500 tokens for your prompt)
- It replies: ~2,000 tokens of analysis
- Cost: $15/M × 50,500 + $75/M × 2,000 = $0.91
This is the heavy case — nearly a dollar for a single request. But analyzing 50-page documents is an outlier. Most daily use is in the pennies range.
Why Agentic Coding Costs More
Agentic coding tools (AI that writes code autonomously) are expensive because they send enormous context windows — often your entire codebase — with every request. A single coding session can consume 100,000+ tokens per request, and those requests happen frequently.
ZDR Chat is designed for conversational use, not agentic coding. You’re in control of what you send. The status bar shows your running cost so you can keep an eye on it.
What $1 Looks Like by Model Tier
| Model | Messages (short) | Pages Analyzed | Code Reviews | Hours of Daily Use |
|---|---|---|---|---|
| Free | Unlimited | Unlimited | Unlimited | Unlimited |
| Budget (GPT-4o Mini) | ~10,000 | ~500 | ~11,000 | ~200+ hours |
| Mid-range (Claude Sonnet) | ~500 | ~25 | ~300 | ~10 hours |
| Premium (Claude Opus) | ~60 | ~1 | ~60 | ~1–2 hours |
No Monthly Subscription Math
Here’s the key insight: you never have to guess whether you’ll use enough to justify a subscription.
- With a $20/month ChatGPT subscription, you’re paying whether you use it or not
- With OpenRouter + ZDR Chat, $1 can last a light user a month
- Heavy users might spend $5–$10/month — and they get access to any model, not just one provider’s
Start With $1
OpenRouter lets you add as little as $1 in credits. Try it for a week. Check the status bar at the end of each session. We think you’ll be surprised how far it goes.