Skip to main content

What does a conversation cost?

Models charge per token, a chunk of text about four characters long. You can paste your own text below to see what it would cost on every model.

Cost calculator

0 tokens (estimated)

What you'd send the model: a question, a document, a prompt.

197 tokens (estimated)

A stand-in for the model's reply. Edit it to match how long you expect answers to be.

Thinking effort (reasoning models)

Reasoning models think step by step before answering, and providers bill those hidden thinking tokens at the same rate as the reply itself. Higher effort means better answers on hard problems, and more tokens on the bill.

Token counts use OpenAI's o200k tokenizer. Other providers split text slightly differently, so treat their counts as close approximations.

What this conversation costs, model by model

Grok 4.1 Fast
xAI
$0.00$0.0003$0.0003
Gemini 3.5 Flash-Lite
Google
$0.00$0.0015$0.0015
Muse Spark 1.1
Meta
$0.00$0.0025$0.0025
Claude Haiku 4.5
Anthropic
$0.00$0.003$0.003
GPT-5.6 Luna
OpenAI
$0.00$0.0035$0.0035
Grok 4.5
xAI
$0.00$0.0035$0.0035
Gemini 3.6 Flash
Google
$0.00$0.0044$0.0044
Gemini 3.5 Flash
Google
$0.00$0.0053$0.0053
Claude Sonnet 5
Anthropic
$0.00$0.0059$0.0059
Gemini 3.1 Pro
Google
$0.00$0.0071$0.0071
GPT-5.6 Terra
OpenAI
$0.00$0.0089$0.0089
Kimi K3
Moonshot AI
$0.00$0.0089$0.0089
Claude Opus 5
Anthropic
$0.00$0.0148$0.0148
Claude Opus 4.8
Anthropic
$0.00$0.0148$0.0148
GPT-5.6 Sol
OpenAI
$0.00$0.0177$0.0177
Claude Fable 5
Anthropic
$0.00$0.0295$0.0295

Chart ready to load.

Not shown (Inkling, Mistral Medium 3.5, GLM-5.2, DeepSeek V4 Pro, Qwen 3.6, Llama 4 Maverick, Llama 4 Scout): open-source models are free to download and run on your own hardware, so there's no fixed per-token price to calculate. Hosted-API pricing for them varies by vendor.

Price per million tokens

Output tokens cost several times more than input tokens on every model. Open-source models aren't shown. They're free to download, so there's no fixed per-token price.

Chart ready to load.