跳到主内容
@wquguru
精选82SemiAnalysis(RSS)行业动态

SemiAnalysis深度解析AI订阅计划经济模型与计费机制

Anthropic Subscriptions Offer 5x+ More Value Than OpenAI

原文
发到 X

Subscription plans are still the primary way consumers and small businesses pay for AI. These plans are highly subsidized—as we previously explained in June—but can still make economic sense as powerful customer acquisition and marketing tools. For example, the goodwill engendered by OpenAI’s generous resets is partially responsible for the recent surge in Codex adoption and has forced Anthropic to repeatedly walk back planned subscription nerfs to avoid getting clobbered in the court of public opinion.

订阅计划仍然是消费者和小企业为 AI 付费的主要方式。这些计划享有高额补贴——正如我们此前在六月所解释的那样——但作为强大的客户获取和营销工具,它们在经济上仍然具有合理性。例如,OpenAI 慷慨的额度重置所赢得的良好声誉,部分促成了近期 Codex 采用率的激增,并迫使 Anthropic 多次撤回计划的订阅限制削减,以避免在舆论场上遭受重创。

Furthermore, because subscription plans are so heavily subsidized, they can also have a large impact on margins and revenue per MW despite only being a small portion of total revenue. Consider the following rough numbers for Anthropic.

此外,由于订阅计划受到高度补贴,尽管仅占总收入的一小部分,它们仍会对利润率以及每兆瓦(MW)的收入产生重大影响。以下以 Anthropic 为例进行粗略估算。

Source: SemiAnalysis Tokenomics Model

来源:SemiAnalysis 代币经济学模型

Despite being just 10% of overall revenue, subscriptions can take up over 40% of inference compute and lower blended revenue per MW by ~$36M. Subscriptions are even more important for OpenAI, as they make up a larger portion of their total revenue. For specific numbers, see our Tokenomics Model.

尽管订阅仅占整体收入的 10%,但它们可能占据超过 40% 的推理算力,并使加权平均每兆瓦收入降低约 3600 万美元。对于 OpenAI 而言,订阅计划更为重要,因为它们在其总收入中占比更大。具体数据请参阅我们的代币经济学模型。

In other words, if you want to accurately model AI lab financials, you need to understand their subscription limits.

换句话说,如果你想准确建模 AI 实验室的财务状况,就必须了解它们的订阅限额。

So how do limits actually work? The way to think about subscriptions is that your monthly payment grants you some numbers of “credits”. Each (model, token type) combo consumes a different amount of credits. Because credit cost ratios can differ dramatically from API price ratios, the “value” of the same plan changes depending on what model and workload you’re running. Put differently, it doesn’t make sense to say a plan is worth $X in isolation—you need to consider the full (plan, model, workload) tuple.

那么限额究竟是如何运作的呢?理解订阅的方式是:你的月度付款会授予你一定数量的“积分”。每种(模型、令牌类型)组合消耗的积分数额不同。由于积分成本比率可能与 API 价格比率大相径庭,因此同一计划的“价值”会根据你运行的模型和工作负载而变化。换言之,孤立地说一个计划值 X 美元是没有意义的——你需要考虑完整的(计划、模型、工作负载)元组。

Here’s how the API-equivalent value of the same $200/month Claude plan changes depending on the model and workload.

以下是相同的每月 200 美元 Claude 计划在不同模型和工作负载下的 API 等效价值变化。

Source: SemiAnalysis Tokenomics Model

来源:SemiAnalysis 代币经济学模型

A single static snapshot isn’t enough either. Labs publicly change their limits all the time with promos and new model releases. They can also silently change limits whenever they want by tweaking credit costs. Ideally, you’d re-check the cost of each (plan, model, token type) daily so you can surface any changes in real-time.

单一的静态快照也不够。实验室经常通过促销和新模型发布公开更改其限额。它们还可以通过调整积分成本随时悄无声息地更改限额。理想情况下,你应该每天重新检查每个(计划、模型、令牌类型)的成本,以便实时发现任何变化。

This is exactly what SemiAnalysis has done with our new Subscriptions Dashboard available exclusively to Tokenomics Model subscribers. Besides every OpenAI and Anthropic subscription, we also track Meta, SpaceXAI, Cursor, Cognition, Z.ai, MiniMax, and Moonshot.

这正是 SemiAnalysis 通过我们专为代币经济学模型订阅者提供的新订阅仪表板所做的。除了所有 OpenAI 和 Anthropic 的订阅外,我们还追踪 Meta、SpaceXAI、Cursor、Cognition、Z.ai、MiniMax 和 Moonshot。

A screenshot of our dashboard showing a small subset of the available data. Source: SemiAnalysis Tokenomics Model

我们仪表盘的截图,展示了可用数据的一个小样本。来源:SemiAnalysis Tokenomics Model

New providers, plans, and models will be added as soon as they’re released. The rest of the article will give an overview of everyone’s current limits. All token and dollar amounts below assume agentic usage unless otherwise specified.

新的提供商、套餐和模型将在发布后尽快添加。文章其余部分将对所有人的当前限制进行概述。除非另有说明,以下所有代币和美元金额均假设用于代理(agentic)使用。

Methodology

方法论

Tokens are generally priced per MTok (million tokens) across the following types:

代币通常按 MTok(百万代币)定价,涵盖以下类型:

  • Input: Fresh tokens added to the LLM’s context window that aren’t cached
  • Cache write: Input tokens that are cached for multi-turn conversations. Generally slightly more expensive than regular input tokens.
  • Cache read: Tokens from previous turns that are already cached in the conversation. Generally extremely cheap compared to uncached tokens.
  • Output: Tokens generated by the model. The most expensive token type.
  • 输入(Input):添加到 LLM 上下文窗口且未缓存的新代币。
  • 缓存写入(Cache write):在多轮对话中缓存的输入代币。通常比常规输入代币稍贵。
  • 缓存读取(Cache read):已在对话中缓存的前几轮代币。与未缓存的代币相比,通常极其便宜。
  • 输出(Output):由模型生成的代币。这是最昂贵的代币类型。

Additionally, some models charge more for tokens above a certain context window. Those models typically compact before that expensive context window is reached, so we also keep our measurements below it.

此外,某些模型对超过特定上下文窗口的代币收取更高费用。这些模型通常在达到昂贵上下文窗口之前进行压缩,因此我们的测量值也保持在窗口之下。

Source: OpenAI

来源:OpenAI

Subscription plans don’t expose this fine grained pricing. Instead, they provide a simple 0 to 100% usage meter across 5-hour and 7-day windows, and sometimes a separate meter for a model like Fable. Subscription tiers are then differentiated by usage multipliers relative only to the provider’s other plans. OpenAI for example used to advertise “Expanded Codex usage” in their $20 Plus plan, “5x more usage than Plus” in the $100 Pro plan, and “20x more usage” in the $200 Pro plan until they cut their $200 plan usage in half and removed all relative usage from their pricing page.

订阅套餐不暴露这种细粒度的定价。相反,它们提供在 5 小时和 7 天窗口内从 0 到 100% 的使用量计量表,有时还为 Fable 等模型提供单独的计量表。订阅层级随后仅根据相对于提供商其他套餐的使用量乘数进行区分。例如,OpenAI 曾在其 20 美元的 Plus 套餐中宣传“扩展 Codex 使用量”,在 100 美元的 Pro 套餐中宣传“比 Plus 多 5 倍的使用量”,在 200 美元的 Pro 套餐中宣传“多 20 倍的使用量”,直到他们将 200 美元套餐的使用量减半,并从其定价页面中移除了所有相对使用量信息。

Screenshot of Claude Code meters. Source: Claude Code CLI

Claude Code 计量表截图。来源:Claude Code CLI

Setup

设置

We compute the subscription-rate of each (plan, model, token type) triple by running experiments that isolate one token type at a time and watching how far the model provider’s meter moves. An experiment is some number of repeated calls using a specific prompt that maximizes one token type while minimizing all the others.

我们通过运行实验来计算每个(套餐、模型、代币类型)三元组的订阅费率,每次实验只隔离一种代币类型,并观察模型提供商的计量表移动了多少。实验是使用特定提示词进行的若干次重复调用,该提示词旨在最大化一种代币类型,同时最小化所有其他代币类型。

Source: SemiAnalysis Tokenomics Model

来源:SemiAnalysis Tokenomics Model

Input, cache writes, and cache reads share the same prompt template. We use a portion of War and Peace since some models will refuse to respond if given large blocks of gibberish. For the input token experiments, we use a random tag on every call to ensure nothing gets cached. The cache write experiments run the same way but with the prompt marked for caching, so each new tag forces a new cache entry. The cache read experiments use a fixed tag, so the first call writes the cache and every repeat reads it.

输入、缓存写入和缓存读取共享相同的提示模板。我们使用《战争与和平》的部分内容,因为如果提供大段乱码,某些模型会拒绝响应。对于输入 token 实验,我们在每次调用时都使用随机标签,以确保没有任何内容被缓存。缓存写入实验以相同的方式运行,但提示标记为可缓存,因此每个新标签都会强制创建一个新的缓存条目。缓存读取实验使用固定标签,因此第一次调用写入缓存,而后续重复调用则读取它。

For the output token experiments, we used a technical essay to force long outputs because models will refuse mechanical prompts like “repeat SemiAnalysis 100,000 times”.

对于输出 token 实验,我们使用了一篇技术文章来强制生成长输出,因为模型会拒绝诸如“重复 SemiAnalysis 100,000 次”这类机械式提示。

Computing the Results

计算结果

For every call we record two things: how many tokens of each type the provider billed, and what the usage meter read. Turning those into a price takes three steps.

对于每次调用,我们记录两件事:提供商对每种类型 token 的计费数量,以及用量计量器的读数。将这些数据转换为价格需要三个步骤。

1. Measure the rate of each token type

1. 测量每种 token 类型的费率

Providers generally report usage on a meter that moves in fixed amounts. This could be a whole percentage point, a credit, a cent etc... A single request often does not move the meter at all, so the cost of one request cannot be measured directly.

提供商通常在一个按固定幅度跳动的计量器上报告用量。这可能是整个百分点、一个积分、一分钱等……单个请求往往根本不会使计量器跳动,因此无法直接测量单次请求的成本。

Instead, we measure in steps. As requests run, we keep a running total of the tokens used. Each time the meter rises, we store that total. The tokens used between two meter moves make up one step, which is the cost of moving the meter by one unit.

相反,我们以步长为单位进行测量。随着请求的运行,我们持续累加已使用的 token 总数。每当计量器跳动一次,我们就存储该总数。两次计量器跳动之间所使用的 token 构成一个步长,即计量器移动一个单位所需的成本。

We drop two partial steps. When a run starts, the meter is already part of the way to its next move, so the tokens before the first move are less than a full step and would make the rate look higher than it is. The tokens after the last move never finish a step, so we drop those as well.

我们丢弃两个不完整的步长。当运行开始时,计量器已经处于向下一次跳动前进的过程中,因此第一次跳动之前的 token 不足一个完整步长,会使费率看起来比实际更高。最后一次跳动之后的 token 未能完成一个步长,因此我们也将其丢弃。

To get a rate, we add up how far the meter moved over the complete steps and divide by the tokens used in them. Providers charge differently for input, output, and cached tokens, so we calculate a separate rate for each.

为了获得费率,我们将完整步长内计量器移动的总幅度除以其中使用的 token 数量。提供商对输入、输出和缓存 token 的收费方式不同,因此我们为每种类型分别计算费率。

Source: SemiAnalysis Tokenomics Model

来源:SemiAnalysis Tokenomics Model

Caveats:

注意事项:

  • Because we see the meter only once per request, a single step can be off by up to one request. These errors don’t add up with more consecutive steps, so we track a range instead of a single number, and the range gets narrower the more steps we count. We keep adding steps until the range is within ±5%, then report the rate across all of them.
  • No request contains only one type of token. Some plans, for example, require a set of instructions on every request. We subtract that extra amount using the prices we measured for the other types.
  • Reading from the cache costs almost nothing on some plans and may not move the meter at all. If the meter does not move after 500M cache read tokens, we assume that cache reads are free.
  • 由于我们对每个请求只能看到一次计量器读数,单个步长的误差可能高达一个请求的量。这些误差在连续多个步长中不会累积,因此我们追踪的是一个范围而非单一数值,且随着计数的步长增多,该范围会逐渐收窄。我们持续增加步长,直到范围缩小至 ±5% 以内,然后报告所有步长的综合费率。
  • 没有哪个请求仅包含一种类型的 token。例如,某些计划要求在每个请求中包含一组指令。我们通过从其他类型测得的单价中扣除这部分额外费用来进行调整。
  • 在某些套餐中,从缓存读取的成本几乎可以忽略不计,甚至可能完全不会消耗用量额度。如果在 500M 缓存读取 token 后用量未增加,我们假设缓存读取是免费的。

2. Convert rates into tokens per window and per month

2. 将费率转换为每个窗口和每月的 token 数量

We read every meter a plan shows after every request. This could be a 5-hour limit, a weekly limit, and for models like Fable, its own weekly limit. So each token type gets a price on each meter. This is how much of that limit a million tokens uses. From each price we work out how many tokens fit e.g. if a million tokens use 5% of a limit, the whole limit holds 20 million.

我们会在每次请求后读取套餐显示的所有用量指标。这可能包括一个 5 小时限制、一个每周限制,以及对于 Fable 等模型特有的每周限制。因此,每种 token 类型在每个指标上都有一个价格。这表示一百万个 token 会消耗该限制的多少比例。根据每个价格,我们可以计算出整个限制能容纳多少 token:例如,如果一百万个 token 消耗了限制的 5%,那么整个限制可容纳 2000 万个 token。

Source: SemiAnalysis Tokenomics Model

来源:SemiAnalysis Tokenomics Model

The 5-hour limit resets many times within a week, so a month is capped by the weekly limits.

5 小时限制在一周内会重置多次,因此月度用量受限于每周限制。

3. Pricing

3. 定价

更进一步:量化金融体系

看懂新闻只是起点——沿量化金融路径,把它变成能交付的工程能力

进入量化体系 →

关联讨论

同一事件的更多信源

相似阅读

关联信息,但可能不是同一事件