GPT-4 pricing and routing profile
GPT-4 from OpenAI lists at $30.00 per million input tokens and $60.00 per million output tokens, with a 8,191-token context window. Prices from OpenRouter's catalogue on 2026-09-28.
Tool callingStructured outputLog probabilities
Cost per 1,000 requests
| Request shape | List cost |
|---|---|
| Classification or routing (600 tokens in, 10 out) | $18.60 |
| Chat reply (2,000 in, 400 out) | $84.00 |
| Retrieval-augmented answer (8,000 in, 500 out) | $270 |
Before prompt caching. Try your own shape in the cost calculator.
Can Gimbix route to it first?
Yes. At least one host return token log probabilities for GPT-4, which is the confidence signal Gimbix uses to let a cheaper model answer first.
Cheaper models that can answer first
Models priced below GPT-4 that expose a confidence signal. Whether they answer your traffic as well is what a Gimbix pilot measures.
| Model | Input / output per 1M | Cheaper on a chat reply |
|---|---|---|
| GPT-4 Turbo | $10.00 / $30.00 | 62% |
| GPT-5 Image | $10.00 / $10.00 | 71% |
| GPT-5.4 Image 2 | $8.00 / $15.00 | 74% |
| GPT-4o (2024-05-13) | $5.00 / $15.00 | 81% |
| Qwen3.8 Max Prime | $4.00 / $12.00 | 85% |
| Ember-1 | $3.00 / $15.00 | 86% |
Questions
How much does GPT-4 cost?
GPT-4 lists at $30.00 per million input tokens and $60.00 per million output tokens on OpenRouter as of 2026-09-28. A typical chat reply of 2,000 input and 400 output tokens costs about $84.00 per 1,000 requests.
What is the context window of GPT-4?
8,191 tokens.
Can a cheaper model replace GPT-4?
Sometimes. Gimbix proves it on your own traffic before routing: the cheaper model must match GPT-4 on one half of your requests and make at most three answers worse on the other.