Compare 2 models, price for price
Choose 2 models. See which costs less for the same tokens and by how much, where one is cheaper to read and the other to write, and what a chat, a support bot or an agent costs a month on each.
Choose 2 models
Per million tokens, 3 read for each 1 written
- GPT-6 AstraOpenAI · $10 in · $50 out · 1.05M window$20
- Grok 4.7xAI · $2 in · $6 out · 500K window$3
Grok 4.7 is cheaper; GPT-6 Astra costs 6.7 times as much for the same tokens.
GPT-6 Astra vs Grok 4.7Moving up from an earlier model
Still on an older model? Choose it, and see what each of its maker's newest models costs against it for the same tokens. Each one opens the full comparison, with 8 jobs priced on both.
Per million tokens, 3 read for each 1 written
- GPT-4oOpenAI · $2.50 in · $10 out · 128K window$4.375
OpenAI's newest models
Who costs how many times whom
Read across: the model in the row costs this many times the model in the column, for the same tokens, 3 read for each 1 written. Yellow is dearer, green cheaper. Each square opens that comparison.
Why the price per token is only the start
Every maker prints 2 main prices: one for the tokens a model reads (input) and one for the tokens it writes (output). On most current models writing costs 3 to 5 times as much as reading, so which model is cheaper can depend on the job. A job that reads a long document and writes a line pays mostly the input price; a chat that writes paragraphs pays mostly the output price. That is why each comparison prices 8 jobs and not one figure.
The same text is not the same number of tokens on every model. Each maker cuts text its own way, and Anthropic says its models from Claude 4.7 on produce about 30 percent more tokens for the same text than earlier ones. A model with the lower price per token can still send the larger bill, so each comparison counts one of our lessons the way both makers count it.
Discounts change the order too. Work that can wait costs half on a batch at the makers that offer one, and text read again from a cache costs a quarter of the input price or less on every current model here. The price table on each comparison shows who offers what.
Thinking is billed as output, and models differ in how much they think and how long they answer. A price list cannot show that. Before you switch, run the same 10 tasks on both models and compare the bills.
Questions
- How do I compare the price of 2 language models?
- Put the 2 price lists side by side, then price a job you actually run, because the input price decides jobs that read a lot and the output price decides jobs that write a lot. Then count the same text with both makers' tokenizers, since the token counts differ. Each comparison here does all 3.
- What is the 3 to 1 blend price?
- One figure for ranking models: the price of 1 million tokens when 3 are read for each 1 written, that is 3 times the input price plus the output price, divided by 4. It suits chat and everyday work. For a job that reads far more than it writes, compare the input price alone.
- Which of each maker's top models costs least?
- Of the top models we list, Grok 4.7 from xAI has the lowest blend price and Claude Fable 5.1 from Anthropic the highest, at 6.7 times as much.
- Can the model with the lower price per token send the larger bill?
- Yes. It may cut the same text into more tokens, think longer before it answers, or write longer answers, and each of those tokens is paid for. Count your own text with the token counter, then compare the bills on a few real runs.