Agent Upstart

Cool things people are building with agents — and practical help getting the most out of them.

Coding playbooks · Reference table

Plan allowances: what each agent plan lets you do

Written November 6, 2026

Messages, credits, and limits for the plans behind Claude, ChatGPT, Copilot, and more — each row linked to the vendor's own page.

Agent allowances live scattered across vendor plan pages and change without a changelog. This table puts them in one place. Every row links to the vendor page the allowance came from. Prices are in US dollars per month unless stated otherwise. Allowances move often, so open the vendor page before buying.

How to read this table

Vendors measure usage differently, and the units do not convert between them. Claude counts usage against rolling 5-hour and weekly windows with no fixed message count. GitHub Copilot counts AI credits, where 1 credit is $0.01, and code completions do not use credits at all. Google counts compute-based Gemini limits that refresh every 5 hours up to a weekly cap, with no fixed message count published. X and Cursor share one weekly usage pool. A bigger number in one vendor's unit is not more than a smaller number in another's. The table shows allowances, not a ranking.

The table

PlanPriceAllowanceSource
Claude Free$0Limited usage every 5 hours and every week. No fixed message count published.Claude pricing
Claude Pro$20, or $17 billed annually ($200 up front)More usage than Free every 5 hours and every week. Claude Code included; chat and Claude Code draw from one pool. Fable paid with usage credits.Claude pricing
Claude Max$1005x Pro usage every 5 hours and every week. $100 a month in API credits. Fable included, up to half the weekly limit.Claude pricing
Claude Ultra$200No 5-hour limit for verified accounts. 10x Pro usage every week. $200 a month in API credits. Fable included, up to half the weekly limit.Claude pricing
Claude Enterprise$20 per seat, plus usage at API ratesBilled annually. Everything in Team plus admin controls, including SCIM, audit logs, and custom data retention.Claude pricing
Claude DuetIncluded with Pro, Max, and Ultra at no extra costResearch preview. Uses the plan's usage limits, including the 5-hour and weekly limits, at the same rate as chats. Extra limit resets are added during the preview; unused resets expire December 15, 2026.Get started with Claude Duet — Anthropic
ChatGPT Free$0Unlimited text chats, subject to abuse guardrails. Limited messages with uploads, image creation, voice chats, deep research, and Codex access.ChatGPT pricing
ChatGPT Plus$20Higher model limits than Free, with expanded messages, uploads, deep research, memory, context, and Codex usage. API use is billed separately and is not included.What is ChatGPT Plus? — OpenAI
ChatGPT Pro 100$100Pro reasoning powered by GPT-6 Astra, longer Codex and ChatGPT Work sessions, maximum deep research, memory, and context. Monthly API credits included. Ultrafast not included.About ChatGPT Pro tiers — OpenAI
ChatGPT Pro 200$200More usage than Pro 100. Weekly GPT-6 Pro message limits; the lower limit was raised to match the higher one. Monthly API credits included. Ultrafast not included.About ChatGPT Pro tiers — OpenAI
ChatGPT Pro 500$500The highest included usage of the three Pro plans. Ultrafast included for GPT-6 Astra and GPT-6.1 Sol: included usage first at 8x the Standard rate, then purchased credits at 6x the Standard credit rate.About ChatGPT Pro tiers — OpenAI
ChatGPT Business, Standard seat$25 per user, or $20 billed annuallyShared workspace with ChatGPT, ChatGPT Work, and Codex. At least 2 paid seats per workspace; at most 200.ChatGPT Business: General FAQ — OpenAI
ChatGPT Business, Premium seat$125 per user, or $100 billed annually5x more usage than a Standard seat, with no 5-hour limit. Included usage is separate from the workspace credit pool.ChatGPT Business: General FAQ — OpenAI
GitHub Copilot Free$02,000 code completions a month. Limited chat and agent use, entry-level models, automatic model selection only.GitHub Copilot plans and pricing
GitHub Copilot Pro$101,500 AI credits a month (1,000 base + 500 flex). Unlimited completions. Copilot Fusion included.Plans for GitHub Copilot — GitHub Docs
GitHub Copilot Pro+$397,000 AI credits a month (3,900 base + 3,100 flex). Unlimited completions. Copilot Fusion included.Plans for GitHub Copilot — GitHub Docs
GitHub Copilot Max$10020,000 AI credits a month (10,000 base + 10,000 flex). Unlimited completions. Copilot Fusion included.Plans for GitHub Copilot — GitHub Docs
GitHub Copilot Business$19 per user1,900 AI credits per user a month, pooled across the organization. Usage beyond the pool is billed separately.Plans for GitHub Copilot — GitHub Docs
GitHub Copilot Enterprise$39 per user3,900 AI credits per user a month, pooled across the organization. Usage beyond the pool is billed separately.Plans for GitHub Copilot — GitHub Docs
GitHub Copilot access$0 with an Enterprise Cloud licence; $10 per user as StandaloneNo included credits. Metered usage is billed from the first use. Same feature and model eligibility as Copilot Enterprise.What's changing with GitHub Enterprise and Copilot billing — GitHub Docs
Google AI Plus$4.992x the standard Gemini limits. 400 GB of storage.Manage daily life with a Google AI plan — Google One
Google AI Pro$19.994x the standard Gemini limits. Gemini 4 Argon, expanded Antigravity and Colab limits, 5 TB of storage. Family sharing with up to 5 people.Manage daily life with a Google AI plan — Google One
Google AI Ultra, 5x$99.995x the AI Pro Gemini limits. Deep Think, higher Argon access, Project Genie, 20 TB of storage.Manage daily life with a Google AI plan — Google One
Google AI Ultra, 20x$199.9920x the AI Pro Gemini limits. The same Ultra features at the highest limits, 30 TB of storage.Manage daily life with a Google AI plan — Google One
X / Cursor Free$0Limited agent requests and chat requests. One subscription covers Cursor and Grok.Cursor pricing
X / Cursor Plus$8Shared usage pool across Cursor and Grok. Grok models only; Grok API access included, billed separately at standard rates.Cursor pricing
X / Cursor Premium$304x the Plus usage. Grok Bot, cloud agents and automations, code hosting, and AI code reviews. Third-party models in Cursor at API list price, billed separately.Cursor pricing
X / Cursor Super$10015x the Plus usage. Priority model access at peak times.Cursor pricing
X / Cursor Ultra$20050x the Plus usage. Earliest access to all products and models.Cursor pricing
X / Cursor TeamsFrom $30 per user4x to 50x the Plus usage by seat tier, pooled. Central billing, usage analytics, and admin controls.Cursor pricing
Meta Muse Free$0Free with a usage limit; the amount is not published. When the limit is reached, wait for it to refresh or upgrade. Limited testing; not available in all locations.About Muse subscriptions — Meta
Meta Muse Power$20500M Muse tokens a week.About Muse subscriptions — Meta
Meta Muse Maximum$1003B Muse tokens a week.About Muse subscriptions — Meta

One attribution note: the Google AI dollar prices in this table ($4.99, $19.99, $99.99, $199.99) did not render on the Google One page as opened; they come from secondary reporting of Google's US plan prices. The allowance multiples and storage figures in those rows are from the vendor page.

Copilot Fusion, and how it is billed

Copilot Fusion is GitHub's model router: instead of picking a model, you pick Fusion and an effort level — Standard, High, or Max — and it chooses one or more models and a working pattern for each prompt. GitHub made it generally available on October 28, 2026, on all Copilot plans, in VS Code, the Copilot app, Copilot CLI, and JetBrains IDEs.

There is no separate Fusion fee. Each model Fusion calls is billed at that model's standard rate in AI credits, and because High and Max can call several models for one task, a task on those levels can use more credits than the same task on Standard. Paid plans get 30% off model costs when using Fusion through November 30, 2026.

What changed recently

ChatGPT Pro 200 was the subject of an allowance cut announced in a September 29 email from OpenAI — then the cut was canceled. The help article says the lower Pro 200 allowance doubled to match the higher one, existing subscribers keep the higher allowance, and the price stays at $200 a month. Pro 100 and Pro 500 also now include twice their previous Work and Codex allowance. Plus usage is unchanged.

Two other changes sit behind rows in this table. Copilot Fusion went generally available on October 28, 2026, with the 30% discount on paid plans running through November 30, 2026. And Gemini 4 Argon, which Google AI Pro and Ultra include, doubles in API price on December 1 — see Gemini 4 Argon price doubling on December 1. That is a price change, not an allowance change, and the table does not conflate the two.

What this table does not do

It does not rank plans. Allowances, caps, and routing differ too much between vendors for a cheapest or best claim to mean anything, and a plan that looks generous on one unit can be the expensive one once the work you actually do is counted. It also does not cover API-only pricing — for token rates, see the model price list — or what happens past a limit, which varies from a hard stop to paid overage depending on the vendor. When a row's allowance changes on the vendor page, the row changes with it. For the full picture of what an agent costs beyond the subscription, see what an agent costs and always-on agent entry costs; for staying inside the allowance you already pay for, see stay inside your subscription allowance and the GPT-6.1 Sol ultrafast Pro tiers.

Related tools

Sources

Back to: Coding playbooks hub · Also: Model prices: what the major models cost · What an agent costs · Always-on agent entry costs · The GPT-6.1 Sol ultrafast Pro tiers · Stay inside your subscription allowance