Major upgrade for Claude subscription plans! The Max 20x tier has been upgraded to Ultra, and the 5-hour usage cap has been completely lifted.
Just now, Anthropic officially announced a major upgrade to Claude's top-tier paid subscription plans!
The original Max 5x plan is officially renamed Max, with the official website monthly fee remaining unchanged at 100 USD.
The original top-tier Max 20x plan is officially renamed Ultra, with the official website monthly fee remaining unchanged at 200 USD.
If you are already a subscriber of these two tiers, you don't need to do anything, and the system will automatically and seamlessly migrate your account.
But what's really exciting is the following announcement —
"Starting today, we will remove the 5-hour usage limit for all verified Ultra accounts!"
What does this announcement mean? People who have never been locked in Claude's "timeout penalty zone" can hardly understand that suffocating feeling.
In the past, even if you paid a high price for the top-tier plan, Anthropic would still use the quota that "resets every 5 hours" to strictly restrict your usage.
When you are in deep flow writing code, you are suddenly limited, which is an extremely frustrating experience.
But now, this restriction has been completely lifted!
Boris Cherny and Alex Albert, the fathers of Claude Code, confirmed this news immediately after the announcement.
For most Ultra subscribers, this benefit takes effect right now, starting today! You can work on your projects for a whole day without staring at the countdown timer while writing code.
The quota war between Anthropic and OpenAI has reached a white-hot stage.
Dig into the official pricing page: the plans have been fully upgraded for high-value users!
This adjustment will cover most existing Ultra subscribers starting from the announcement date. Newer accounts will still have temporary restrictions, which will be lifted after automatic verification, and users will be notified via email.
It seems that Anthropic has kept a backdoor to prevent "welfare scalpers" from reselling API access.
Meanwhile, the 100 USD/month Max plan still retains the 5-hour usage cap.
In addition, the total weekly quota remains unchanged. Even so, breaking the 5-hour limit is still a major concession on Claude's path of commercialization!
What is the official's underlying strategy? We immediately checked the brand new pricing page for more details.
After checking, we found that Anthropic has divided users into very clear tiers with extremely precise positioning, comparable to Apple's product pricing strategy led by Tim Cook.
In their view, users are divided into the following tiers.
Tier 1: Free users and light users
Free Plan: Experience Claude 3.5 Sonnet at zero cost, but with very limited quota.
Pro Plan (17 USD/month): The standard configuration for most regular office workers. It offers more usage than the free tier, supports features including Projects and Claude Code, and unlocks Claude's core productivity capabilities.
Tier 2: Hardcore developers (the core of this upgrade)
Max Plan (100 USD/month, originally Max 5x): This is for heavy users who find the Pro plan completely insufficient. At this tier, your computing power quota is greatly increased, with extremely high cost performance.
Ultra Plan (200 USD/month, originally Max 20x): This is the protagonist of today, the exclusive premium Ultra Club! Users who purchase this plan are almost all senior power users.
In addition, the developer plan includes all the rights and interests of the personal plan, plus more additional benefits, while the price remains unchanged!
The logic behind Anthropic's removal of the 5-hour limit for Ultra is very clear —
Since power users are willing to pay 200 USD, we must provide them with real "first-class service".
High-value users should not feel restricted or dissatisfied after paying for premium services. Anthropic's subscription system has finally become user-centric.
Hidden Easter Egg
Now, Anthropic has fully sorted out the rights and interests of all its paid subscription plans.
However, after the revision, some users found that after Max 20x was renamed to Ultra, the official website stated that "the weekly quota is 10 times that of Pro"?
The left side is the new official pricing page, which says "10x Pro weekly usage"; the right side is the pop-up window when users enter the account backend to switch plans, which still retains the old copy of "20x Pro usage".
Is Anthropic falsely advertising and secretly cutting the quota?
You will understand after checking the official documents carefully.
Let's first clarify what the 20x quota actually means.
The "20x" advertised by the old Max 20x only refers to the message volume that is 20 times that of Pro users within a single 5-hour window. When extended to 7 days, the total weekly upper limit of the 200 USD tier has always been fixed at 10 times that of Pro.
After Ultra cancels the single 5-hour limit, the "20x in single window" loses its reference, so the official website uniformly uses the description "the total weekly upper limit is 10 times that of Pro". As for the remaining 20x text in the backend pop-up window, it is purely an oversight where the old pop-up was missed during frontend updates.
Heavy users should pay attention to a hidden rule.
If you frequently call the top-tier Fable model under the Ultra plan, there is an independent weekly restriction: you can use up to 50% of your total weekly quota for this model, which is equivalent to 5 times the weekly quota of the Pro plan.
In addition to keeping the total weekly quota completely unchanged, Anthropic has also given Ultra subscribers a substantial big easter egg: the monthly free API quota for the Claude platform (formerly Anthropic Console) has been doubled from 100 USD under the Max tier to 200 USD!
This 200 USD free API credit is automatically credited in each monthly billing cycle and will be cleared at the end of the month. You can use it to run automated scripts via API, or experience the Claude Managed Agents service, all for free.
Since the pricing rules of AI giants' plans are so complicated these days, even David Leen, the head of Netflix platform, complained: I'm afraid we need to use Lean to figure out all these discounts.
Wait, this looks a little familiar.
Both priced at 200 USD? The two giants launched a fierce overnight battle
The timing of this upgrade is very coincidental.
Just yesterday, OpenAI just officially announced: the planned quota cut is canceled, and the usage quota for all tiers of ChatGPT Pro is doubled.
The three premium tiers Pro 100, Pro 200, Pro 500, and Business Premium, all usage caps are doubled, effective immediately!
The original planned quota cut for Pro 200 has also been quickly removed.
The reason is that OpenAI previously thought that its computing power could not support the load, so it originally planned to secretly reduce the quota of Pro 200, and cut the usage limit for Work and Codex.
As soon as the news came out, the entire community was furious: "Refund our money! Refund immediately!"
This public outcry shocked OpenAI's management. The policy was adjusted from cutting the quota by half to doubling it, all within one night.
Tibo has been temporarily crowned as "Saint T"
Why are both companies now targeting these expensive plans priced at 100 USD and 200 USD?
Obviously, the 20 USD/month Plus and Pro users are no longer the priority target for large model companies.
What they are competing for now are high-value customers: professional programmers, analysts, and architects.
Most heavy users of Claude are not sensitive to prices, what they want is smooth experience and stability.
The removal of the 5-hour limit is a major positive update.
For Pro 200 and Pro 500 users of OpenAI, driving them away due to trivial computing power costs will be a huge blow to OpenAI's ecosystem.
You lift the usage limit today, and I double the quota tomorrow, which reflects this core truth —
The gap between underlying large models is narrowing, and computing power quota and user experience are becoming the real competitive moat.
References:
https://claude.com/pricing#developer
https://x.com/Joshua_WD/status/2115620634386419984