Update, 1 September 2026. This did not happen: Anthropic cancelled the increase, $2/$10 is now the standard price, and every dollar figure in this post stands unchanged. Detail in my reply below. The original comment is kept as written.
One caveat on the dollar figures above, since it will bite in a few weeks.
Claude Sonnet 5's $2/$10 is introductory pricing and reverts to $3/$15 on 1 September. Every dollar amount in this post was computed at the old rate, so $0.351 per run and $10,534 a month will both move. I will update the post and the calculator that day.
The token counts do not change. 21,700 raw against 135,600 billed, and the 6.2x multiplier, are properties of the loop rather than the price. That is sort of the point: the multiplier is what your architecture costs you, and the rate card only scales it.
If you spot a stale price before I do, the data is in models.json in the repo with a source URL and a verification date on every rate. An issue with the provider's pricing link is the fastest way to get it fixed.
Following up on my own caveat above, on the day it was meant to bite: the increase did not happen.
Anthropic cancelled it. Their pricing page now carries a note saying the $2/$10 announced at launch as introductory through 31 August is the standard price, and that the scheduled move to $3/$15 "will not occur". So every dollar figure in this post stands unchanged. $0.351 per run, $10,534 a month, $0.109 cached, the 20-step comparison, all of it.
I did find one thing worth correcting while I was in there, and it cuts against my own numbers. Anthropic run two tokenizers. Claude 4.7 and later, which covers Sonnet 5, produce roughly 30% more tokens for the same text than Haiku 4.5 does. The table in this post compares them at a single token count, which quietly assumes the same text costs the same number of tokens on both. It does not. Switching to Haiku saves somewhat more than the 50% shown, because you are also billed for fewer tokens to say the same thing.
The calculator's limitations section says this now. I have not tried to model it, because the size of the gap depends on your actual text and I would rather show you a number I can source than one I estimated.
For further actions, you may consider blocking this person and/or reporting abuse
We're a place where coders share, stay up-to-date and grow their careers.
Update, 1 September 2026. This did not happen: Anthropic cancelled the increase, $2/$10 is now the standard price, and every dollar figure in this post stands unchanged. Detail in my reply below. The original comment is kept as written.
One caveat on the dollar figures above, since it will bite in a few weeks.
Claude Sonnet 5's $2/$10 is introductory pricing and reverts to $3/$15 on 1 September. Every dollar amount in this post was computed at the old rate, so $0.351 per run and $10,534 a month will both move. I will update the post and the calculator that day.
The token counts do not change. 21,700 raw against 135,600 billed, and the 6.2x multiplier, are properties of the loop rather than the price. That is sort of the point: the multiplier is what your architecture costs you, and the rate card only scales it.
If you spot a stale price before I do, the data is in models.json in the repo with a source URL and a verification date on every rate. An issue with the provider's pricing link is the fastest way to get it fixed.
Following up on my own caveat above, on the day it was meant to bite: the increase did not happen.
Anthropic cancelled it. Their pricing page now carries a note saying the $2/$10 announced at launch as introductory through 31 August is the standard price, and that the scheduled move to $3/$15 "will not occur". So every dollar figure in this post stands unchanged. $0.351 per run, $10,534 a month, $0.109 cached, the 20-step comparison, all of it.
Source, checked today: platform.claude.com/docs/en/about-claude/pricing
I did find one thing worth correcting while I was in there, and it cuts against my own numbers. Anthropic run two tokenizers. Claude 4.7 and later, which covers Sonnet 5, produce roughly 30% more tokens for the same text than Haiku 4.5 does. The table in this post compares them at a single token count, which quietly assumes the same text costs the same number of tokens on both. It does not. Switching to Haiku saves somewhat more than the 50% shown, because you are also billed for fewer tokens to say the same thing.
The calculator's limitations section says this now. I have not tried to model it, because the size of the gap depends on your actual text and I would rather show you a number I can source than one I estimated.