SemiAnalysis tests show Claude subscription quotas quietly cut in gray test, AI membership shrinkage sparks controversy

Claude Caught Quietly Cutting Quotas! AI Memberships Are Collectively Turning into "Rigged Scales"

Ifanr Download the app for exclusive early access October 23, 2026 15:13:49 From Beijing

Recently, quite a few users have complained that Claude's quota consumption speed has suddenly become extremely abnormal.

Just when everyone was wondering whether they were experiencing a collective hallucination, the well-known semiconductor and AI research firm SemiAnalysis dropped a hard piece of evidence.

They posted on X to sound the alarm, confirming that users' feelings were not wrong at all: Anthropic is indeed quietly running an A/B gray test on Claude's subscription quotas in the background.

On the same day, SemiAnalysis ran an Opus 5.5 Agent workload test using two completely identical $200 Max 20x accounts, and the results revealed a very clear discrepancy.

The test group's account B was given a 5-hour window quota of 2.076 billion tokens, 19% higher than the normal account A's 1.74 billion tokens.

But when you look at the more critical weekly total ammunition, account B shrank from the normal 6.579 billion tokens to 5.266 billion tokens, wiping out a full 20%.

Converted into monthly API-equivalent value, the normal account A can generate $11,726 worth of compute value per month, equivalent to 59 times the $200 monthly fee.

Meanwhile, the unlucky account B shoved into the gray test group saw its monthly equivalent value drop to $9,385, with the subsidy multiplier falling to 47 times the monthly fee—heavy users are having a solid 20% of their quota taken away every month.

No wonder these past couple of days countless Vibe Coding folks who paid for the $100 and $200 plans suddenly discovered that the Opus 5.5 in their hands had completely changed flavor.

People used to joke that ChatGPT's catchphrase was "You make a valid point," but now with Claude Code, its catchphrase has directly become "You've reached your usage limit."

Another X user reported that last week they could barely feel any limits, but today, before they had even run any heavy tasks, 23% of their quota was already gone.

Another user named Priya was even more helpless: their daily work had just gotten started, and the 90% progress bar was already bottoming out.

What's even more interesting is the official damage-control posture right from the start.

Facing doubts from across the internet, Claude Code technical staff member Lydia Hallie came out to reassure everyone, explicitly stating that there had been no adjustment to quotas and suggesting that users who felt consumption was fast should run the feedback command and DM her their ID for investigation.

Right after that, she offered a set of blame-shifting deductions, asking users whether they frequently left in the middle of long conversations for more than 1 hour, causing the context cache to expire, or whether they had opened too many sub-Agents in a session.

Claude Code's core developer Boris Cherny has also just stated that the team is urgently investigating the cause and will give everyone an explanation soon.

The most unbearable part of this whole affair is that just half a month ago, SemiAnalysis had just given Anthropic a highly valuable good-guy card.

At the time, in order to dig out the real padding behind each company's plans, the analysts wrote a test program that repeatedly stuffed the text of the classic novel War and Peace into major models, specifically watching that blind-box progress bar that only shows 0% to 100% but never indicates the specific number of tokens.

After a round of extreme stress testing and calculation, among plans in the same $200 tier, the API-equivalent value provided by the Claude subscription using Opus 5.5 was about 5 times that of OpenAI's same-tier plan using GPT-6.1 Sol.

The direct trigger for the gap widening to a runaway 5 times was that neighboring OpenAI struck earlier and with even less restraint.

OpenAI had previously halved the token quotas for each model tier of its $200 Pro plan on the spot, and even erased the relative-multiplier promotional language from its official pricing page.

After cutting the $200 plan, OpenAI turned around and rolled out a new $500-per-month tier.

But actual testing showed that for the extra $300, the top-tier Astra model's quota was only about 21% more than the pre-adjustment $200 plan—the real gimmick had become the ultra-fast output mode of 300 tokens per second.

In short, the hundreds of extra dollars users paid did not buy much additional quota; a large part of it was simply paying for output speed. Now, the seemingly upright Anthropic has ultimately taken the same approach.

The reason may be that, in the eyes of vendors such as OpenAI and Anthropic, the heavily subsidized monthly subscription system has become the biggest drag on profit margins.

According to SemiAnalysis's calculations, the subscription business contributed only about 10% of Anthropic's total revenue but consumed more than 40% of its inference compute.

But does the large-model business really have to be sustained only by continuously shrinking user quotas to maintain gross margins? Turning our gaze back home, look at the engineering answers that domestic model vendors have just delivered—it is a completely different picture.

Just these past couple of days, DeepSeek published its technical practice for large-scale inference deployment of V4.1 Flash and officially open-sourced its cloud service inference framework DeepInfer. Using common 8-GPU machines for prefill and 64-to-128-GPU supernodes for decoding, a single DeepSeek cluster achieved a daily throughput of 155 trillion tokens.

There are two very hardcore data points in this report.

The first is that DeepSeek supports retaining KVCache context for more than 72 hours at the system design level, solving the cost challenge of long-cycle context reuse with solid engineering optimization.

The second data point directly lays bare the money-making capability of this business: even calculated at off-peak low prices, the overall revenue of DeepSeek's inference system reached $38 per GPU per hour.

Do the simplest math: one GPU can generate more than $900 in revenue per day over 24 hours, fully proving that as long as the underlying inference architecture and scheduling optimization are done well, the AI model business can absolutely achieve extremely high gross margins.

In the mobile internet era, the marginal cost of software and content services is almost zero. After users activate a monthly membership, no matter how frequently they use it, the extra bandwidth cost the platform pays is negligible.

And what disgusts users most may not necessarily be the price increase itself, but the lack of basic decency and transparency in the whole process.

Vendors do not publish specific consumption standards, only giving a vague percentage progress bar; either they secretly degrade intelligence, or they quietly run A/B tests in the background to lower quotas.

Users pay real money for their subscriptions on time every month, yet the service standard they buy can be tampered with at any time without their knowledge. This kind of trust overdraft—treating users like fools—is probably harder to make up for than the consumption of compute.

"Special Statement: The above content (including videos, pictures, or audio if any) is uploaded and published by users of Dafeng Hao, a self-media platform under Phoenix Net. This platform only provides information storage space services.

Notice: The content above (including the videos, pictures and audios if any) is uploaded and posted by the user of Dafeng Hao, which is a social media platform and merely provides information storage space services."