Claude Max vs API cost: $100 plan or tokens?

According to APITariff, Claude Max starts at $100 per month on 2026-10-14 and its 5x tier lists $100 per month in API credits, per claude.com/pricing, so Claude Max vs API cost is a question of what $100 of tokens buys. Counting token dollars alone, API spend under $100 a month costs less than the Max 5x price. At $100 or more, the plan’s credit offsets that spend only if it applies to the model you call, and the pricing page does not say whether it does. At Claude Sonnet 5.5 rates of $2 input and $10 output per 1M tokens, $100 buys 25M tokens at a blended 3 input to 1 output mix. Anthropic publishes no token cap for Max, only “5x or 20x more usage than Pro” per 5-hour session, so a Max plan and an API bill compare in dollars of API usage and in nothing else. Max also adds session limits, higher output limits and priority access, which an API bill has no line for.

Last updated 2026-10-14 · 4 Claude models · sources: claude.com/pricing, Anthropic API pricing page, APITariff daily fetch of OpenRouter and kie.ai.

What does each Claude plan cost next to API credits?

Max is the only plan that lists monthly API credits: $100 on Max 5x and $200 on Max 20x, read from claude.com/pricing on 2026-10-14. The page prints a plan price for Max only as “From $100”.

Plan	Price on claude.com/pricing	Monthly API credits
Free	$0	No
Pro	$20 per month, or $200 billed up front for a year	No
Max 5x	from $100 per month	$100 per month
Max 20x	not printed as a separate figure in our read	$200 per month

Source: claude.com/pricing, raw page read 2026-10-14. The page says Max gives “5x or 20x more usage than Pro” per 5-hour session, higher output limits and priority access at high traffic times. It prints no token count for any plan.

How many tokens do $20, $100 and $200 buy on each Claude model?

At a blended 3 input to 1 output mix, $100 of API spend buys 25M tokens on Claude Sonnet 5.5, 12.5M on Opus 5.5, 5M on Fable 5.1 and 500M on Haiku 5.5, at the 2026-10-14 rates below.

We compared 4 Claude models at 3 spend levels ($20, $100 and $200) and assumed a 3 to 1 mix; change it in the token calculator. The blended rate is (3 x input + output) / 4 per 1M tokens.

Model	Input / output $ per 1M	Blended $ per 1M	$20 of spend buys	$100 buys	$200 buys
Claude Haiku 5.5	0.10 / 0.50	0.20	100M	500M	1,000M
Claude Sonnet 5.5	2.00 / 10.00	4.00	5M	25M	50M
Claude Opus 5.5	4.00 / 20.00	8.00	2.5M	12.5M	25M
Claude Fable 5.1	10.00 / 50.00	20.00	1M	5M	10M

Source: Anthropic API pricing page, standard rates, read 2026-10-14, matched against the OpenRouter rows APITariff fetched at 07:10 UTC the same day. Above 100,000 prompt tokens, Haiku 5.5 rates rise to $0.50 and $2.50; the table uses the lower rate for prompts of 100,000 tokens or fewer. $20 is the Pro monthly price, which includes no API credits.

At how many calls does the API reach $100?

The API reaches $100 at 20,000 calls a month on Claude Sonnet 5.5 and at 4,000 calls on Fable 5.1, with calls of 1,000 input and 300 output tokens.

Counting token dollars only, monthly API spend under $100 costs less than the Max 5x price. At $100 or more the comparison depends on credit terms the page does not state. Sonnet 5.5 at 2,000 calls of that size bills $10.

Model	Cost per call	Calls per month to reach $100
Claude Haiku 5.5	$0.00025	400,000
Claude Sonnet 5.5	$0.005	20,000
Claude Opus 5.5	$0.01	10,000
Claude Fable 5.1	$0.025	4,000

Worked by APITariff from the rates above on 2026-10-14. Cost per call = (1,000 x input + 300 x output) / 1,000,000. No caching, no batch. Anthropic lists batch for Sonnet 5.5 at $1 input and $5 output per 1M, half the standard rate, which doubles the call counts for that model.

What does Max include that an API bill does not?

Max includes usage limits measured against Pro, higher output limits and priority access at high traffic times; Claude Code is listed on Pro as well.

• Usage: Max 5x and 20x give “5x or 20x more usage than Pro” per 5-hour session (claude.com/pricing, 2026-10-14). An API bill has no session meter on the pricing page; rate limits are not covered here.
• Fable: Max lists “50% of weekly limits” for Fable, Pro lists usage credits and Free lists No.
• Claude Code: Pro and Max both list it as Yes. The page says heavy Claude Code sessions can switch to pay-as-you-go API credits through a Console account.

Is the Claude API free with a Claude subscription?

No for Free and Pro, and partly for Max: claude.com/pricing lists “No” for monthly API credits on Free and on Pro, and $100 or $200 per month on Max 5x and Max 20x, read on 2026-10-14. Max does not make API calls free; the credits are a fixed dollar amount. The page does not say whether unused credits roll over or what they apply to beyond “to build on the Claude Platform”, so we list both as not covered.

What do Haiku 5.5, Sonnet 5.5 and Opus 5.5 cost through kie.ai?

kie.ai is a reseller and gives none of the Max plan features. It lists three of the four models at 40% to 60% of Anthropic’s list rate: $0.80 input and $4.00 output per 1M tokens for Claude Sonnet 5.5, against $2 and $10 direct, from the kie.ai rows in APITariff’s daily fetch of 2026-10-14 07:10 UTC.

Model	kie.ai in / out $ per 1M	Direct in / out $ per 1M	Calls to reach $100 via kie.ai
Claude Haiku 5.5	0.06 / 0.30	0.10 / 0.50	666,667
Claude Sonnet 5.5	0.80 / 4.00	2.00 / 10.00	50,000
Claude Opus 5.5	1.60 / 8.00	4.00 / 20.00	25,000
Claude Fable 5.1	not listed	10.00 / 50.00	not available

kie.ai lists the older Claude Fable 5 at $4 input and $20 output, and no row for Fable 5.1 (checked 2026-10-14). kie.ai bills in credits and the dollar figures are its own conversion. No Claude request of ours has gone through kie.ai, and we have not read its retention terms.

Affiliate link. If you sign up through it we may earn a commission. The rates above come from kie.ai’s rows in our daily fetch, not from the link. kie.ai’s pricing page lists the rest.

What is not on this page

• The Max 20x price as its own figure: the page text we read prints only “From $100” for Max.
• Any token limit for Max, Pro or Free: Anthropic prints none.
• Whether Max API credits expire, roll over or apply to every model.
• Quality, speed or benchmark scores: none of the four models has run on our bench.
• Amazon Bedrock, Google Cloud and Microsoft Foundry rates, which we did not fetch, and negotiated volume discounts.

Cache and batch rows for each Claude model sit in the Anthropic API pricing guide, and the same per-token arithmetic across 8 providers is in what 1,000 tokens cost.

This guide is an AI-assisted draft built from APITariff’s own price data, then checked and edited by a person. Its chart is drawn from the rates above, and no image here is AI-generated.

Change log

• 2026-10-14First edition. Plan prices read from claude.com/pricing, token rates from the Anthropic pricing page, both on 2026-10-14.