ChatGPT Plus vs Claude Pro vs Google AI Pro: Complete Comparison and Report on Price, Usage Limits, Models, and Coding Agents
ChatGPT Plus, Claude Pro and Google AI Pro are the three subscriptions most people weigh when they decide to pay for an AI assistant, and all three sit at the same point of their vendors' price lists: $20 a month for ChatGPT Plus and Claude Pro, $19.99 a month for Google AI Pro. OpenAI sells Plus between its $8 Go plan and Pro plans that start at $100. Anthropic sells Claude Pro between its Free plan and the $100 Max and $200 Ultra plans. Google sells AI Pro between the $4.99 Google AI Plus plan and Google AI Ultra, which starts at $99.99. Underneath the matching price, the three plans give access to different flagship models, meter usage in different ways and bundle very different extras.
The price gap is close to zero on monthly billing, and the annual options decide it: Claude Pro costs $200 a year and Google AI Pro $199.99 a year, about $16.67 a month, while OpenAI's pricing pages list no annual price for Plus, so a year of Plus costs $240. On what each plan includes, Google AI Pro carries the largest bundle (5 TB of storage, Gemini in Gmail and Docs, family sharing with up to five people and a 1M-token context window in the Gemini app). ChatGPT Plus is the only one that publishes per-model estimates for its agent and coding allowance, with between 15 and 160 GPT-6.1 Sol messages every five hours in Work and Codex. Claude Pro is the only one that puts its Opus 5.5 model and the Claude Code terminal agent inside the plan allowance together with a context window of up to 1M tokens, at the cost of the least specific usage disclosure.
For a buyer who wants one assistant for writing and research plus office files, Google AI Pro gives the most per dollar. For a developer who wants a coding agent in the terminal under the same subscription, Claude Pro and ChatGPT Plus are the two real candidates, and the choice turns on whether the work needs Opus 5.5 or the model range and published limits of Codex. All prices and limits below were checked on the vendors' official pricing pages and help centers on November 4, 2026.
··········
CHATGPT PLUS, CLAUDE PRO AND GOOGLE AI PRO: PRICES AND BILLING OPTIONS.
OpenAI's ChatGPT Help Center states that "the Plus plan costs $20/month", and the ChatGPT pricing documentation lists the same amount. Neither page shows an annual price for Plus. OpenAI's help article on Plus adds that customers in the EU, the UK and Turkey can get a refund if they cancel within 14 days of purchase, and that subscription fees are otherwise non-refundable.
Anthropic's pricing page lists Claude Pro at $20 billed monthly, or $17 a month on an annual subscription billed up front at $200. Anthropic's help article on the Pro plan says that prices appear in local currency where supported and that some regions include tax in the displayed price while others add it at checkout.
Google's Google One page for AI plans offers Google AI Pro at $19.99 a month or $199.99 a year. The Gemini subscriptions page links to a free-trial offer, but neither page states the length or terms of that trial in the content available this week, so the trial is left out of the comparison below.
........
Plan | Monthly price | Annual price | Effective monthly on annual | Year of monthly billing |
ChatGPT Plus | $20 | Not published | Not available | $240 |
Claude Pro | $20 | $200 | $16.67 (listed as $17) | $240 |
Google AI Pro | $19.99 | $199.99 | $16.67 | $239.88 |
........
The annual plans of Anthropic and Google save $40 and $39.89 a year against monthly billing. On monthly billing the three plans are within one cent of each other.
··········
WHERE EACH $20 PLAN SITS IN ITS VENDOR'S PRICE LADDER.
The three plans are the middle rung of three different ladders, and the rungs above and below them explain much of what each $20 plan leaves out. OpenAI places ChatGPT Go at $8 below Plus and offers Pro at $100, $200 or $500. Anthropic has no plan between Free and Pro, and above Pro it sells Max at $100 and Ultra at $200; Ultra is the renamed Max 20x plan, described in our report on the Claude Ultra rename. Google sells AI Plus at $4.99, then AI Pro, then AI Ultra from $99.99, with a $199.99 Ultra tier that Google describes as offering 20 times the usage limits of AI Pro.
........
Vendor | Plan below | The $20 plan | Plans above |
OpenAI | Go, $8 | ChatGPT Plus, $20 | Pro at $100, $200 or $500 |
Anthropic | Free, $0 | Claude Pro, $20 | Max $100, Ultra $200 |
AI Plus, $4.99 | Google AI Pro, $19.99 | AI Ultra from $99.99 (also $199.99) |
........
Google's ladder has the cheapest paid entry point, and its $4.99 Plus plan already doubles the free usage limits and adds 400 GB of storage. Anthropic's ladder has the largest step from free to paid, which makes Claude Pro the first plan that adds Opus models, Claude Code and Research.
··········
MODELS INCLUDED IN CHATGPT PLUS, CLAUDE PRO AND GOOGLE AI PRO.
The model lists are where OpenAI's own pages disagree with each other. The plan comparison on chatgpt.com still names GPT-5.5 Thinking as the advanced reasoning model of Plus, while OpenAI's ChatGPT documentation for Work and Codex lists GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Sol, GPT-6 Luna and GPT-6 Astra with usage estimates for Plus. OpenAI's help article on GPT-6 Astra says that "Plus includes Astra in Work and Codex, but not GPT-6 Pro in Chat", so the GPT-6 Pro mode in the Chat tab stays reserved for Pro, Business and Enterprise plans.
Anthropic's pricing page lists Opus, Sonnet and Haiku for Pro, against Sonnet and Haiku on the Free plan. The current versions on the Claude Platform model pages are Claude Opus 5.5, released on September 22, 2026, Claude Sonnet 5.5 and Claude Haiku 5.5. Claude Fable 5 and Fable 5.1 are available on Pro but sit outside the plan's usage limits: Anthropic's help article says Pro users pay for them with pay-as-you-go usage credits.
Google's subscriptions page lists "Access to Gemini 4 Argon" for AI Pro, together with the Pro and Flash models, and also names Gemini 3.1 Pro. When an AI Pro subscriber reaches the limit, Google's Gemini Apps help says the app continues with Flash-Lite.
........
Plan | Flagship models listed | Models outside the plan allowance | Fallback at the limit |
ChatGPT Plus | GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Astra, GPT-6 Luna (Work and Codex); GPT-5.5 Thinking on the plan page | GPT-6 Pro in Chat (not on Plus) | Not published; ChatGPT credits extend usage |
Claude Pro | Opus 5.5, Sonnet 5.5, Haiku 5.5 | Fable 5 and Fable 5.1 (usage credits) | Usage credits, if enabled |
Google AI Pro | Gemini 4 Argon, Gemini 3.1 Pro, Flash models | Not stated | Flash-Lite |
........
On model access, the three plans restrict their strongest options in different ways: GPT-6 Astra on Plus is limited to Work and Codex, Claude Pro keeps Fable outside the allowance while including Opus 5.5, and Google lists Gemini 4 Argon for AI Pro without a per-model quota.
··········
HOW USAGE LIMITS WORK ON CHATGPT PLUS, CLAUDE PRO AND GOOGLE AI PRO.
All three vendors now use the same basic structure: a rolling five-hour window plus a weekly cap. The difference is how much each one tells subscribers about the size of those windows.
Anthropic's help article says Pro plans have "a five-hour limit and a weekly usage limit that applies across all models", and its pricing page adds that Pro gives at least five times the usage of the Free plan per five-hour session. Usage is shared across Claude on the web, desktop and mobile apps and Claude Code. Anthropic publishes no message counts for Pro and says the number varies with message length, file length, conversation length and model. Anthropic doubled the five-hour Claude Code limits on Pro in May 2026 and increased five-hour limits again with the Opus 5.5 launch in September, without giving figures in either announcement.
Google's Gemini Apps help states that AI Pro has four times the standard limits, AI Plus twice, and AI Ultra five or twenty times the AI Pro limits depending on the tier. The limit "refreshes every 5 hours until you reach your weekly limit." Google publishes no prompt counts per model.
OpenAI publishes the most detail, but only for Work and Codex. The ChatGPT pricing documentation gives estimated local messages per five hours for each model on Plus and says weekly limits may also apply. For the Chat tab, the plan page says only "expanded messages and uploads". OpenAI's Help Center has described a move to one shared token-based limit for ChatGPT and Codex in December, as reported in our article on the shared ChatGPT and Codex usage limit.
........
Limit rule | ChatGPT Plus | Claude Pro | Google AI Pro |
Five-hour window | Yes (Work and Codex) | Yes | Yes |
Weekly cap | "Weekly limits may also apply" | Yes, across all models | Yes |
Published size | Per-model estimates for Work and Codex | At least 5x Free per session | 4x standard limits |
Shared across products | Work and Codex share one allowance | Chat apps and Claude Code share one allowance | Not stated |
Paid overflow | ChatGPT credits | Usage credits | AI credits for Flow and Antigravity |
........
··········
CHATGPT PLUS WORK AND CODEX ESTIMATES BY MODEL.
OpenAI's ChatGPT pricing documentation gives these ranges for Plus, measured as local messages in a five-hour period. OpenAI calls them estimates, not fixed limits, and says cloud tasks can consume more of the allowance than local messages.
........
Model | Plus, messages per 5 hours | Pro plans |
GPT-6 Astra | 5-45 | 25-225 (Pro 5x), 100-900 (Pro 20x) |
GPT-6.1 Sol | 15-160 | No five-hour limit currently |
GPT-6.1 Terra | 35-285 | No five-hour limit currently |
GPT-6 Sol | 15-150 | No five-hour limit currently |
GPT-6 Luna | 350-3,000 | No five-hour limit currently |
........
The Astra figures for Pro come from OpenAI's help article on managing GPT-6 Astra usage, while the pricing documentation says Pro plans "currently have no five-hour limit" for the other models. The same documentation says Fast mode consumes 2.5 times the standard allowance, that image generation inside Codex uses the allowance three to five times faster than a turn without images, and that the Ultrafast mode is not available on Plus. GPT-6.1 Sol runs with a default 500,000-token context window in Codex and ChatGPT Work, according to OpenAI's model documentation.
The ranges are wide because a short local edit and a long multi-file refactor count as one message each. At the low end, five Astra messages per five hours is enough for occasional hard problems, while the 350 to 3,000 messages of GPT-6 Luna cover high-volume routine work.
··········
CONTEXT WINDOWS AND FILE UPLOAD LIMITS ON THE THREE PLANS.
Context window size is the clearest technical difference between the plans. ChatGPT Plus has the smallest window in chat, Claude Pro and Google AI Pro both advertise up to 1M tokens.
........
Limit | ChatGPT Plus | Claude Pro | Google AI Pro |
Chat context, fast model | 54K tokens (about 40 pages of input) | Up to 1M, varies by model | 1M tokens |
Chat context, reasoning model | 256K tokens (about 320 pages of input) | Up to 1M, varies by model | 1M tokens |
Work and Codex context | 500K default for GPT-6.1 Sol | Shared with chat | Not stated |
File uploads | "Expanded" (no figure on the plan page) | 500 MB per file, 20 files per chat, PDFs up to 1,000 pages | Up to 1,500 pages of files, code up to 30K lines |
Project files | Projects included | 30 MB per file, unlimited count within context | Gemini Notebook up to 300 sources per notebook |
........
The ChatGPT figures come from the context table on OpenAI's pricing page, which lists the same 54K and 256K values for Plus and Go. Anthropic's file limits come from its help article on uploads, which gives one set of limits without plan differences. Google's 1M-token figure for AI Pro and AI Ultra comes from the Gemini Apps help, which lists 32K tokens without a plan and 128K tokens on AI Plus. The 1,500-page and 30,000-line figures appear on Google's subscriptions page without being tied to a specific plan.
For anyone who regularly pastes long contracts, codebases or research papers into a chat, the gap between 256K tokens on Plus and 1M tokens on the other two plans is the most concrete reason to choose Claude Pro or Google AI Pro.
··········
DEEP RESEARCH, AGENTS AND CODING TOOLS IN CHATGPT PLUS, CLAUDE PRO AND GOOGLE AI PRO.
All three plans include a research mode that writes cited reports. ChatGPT Plus gets "expanded deep research and agent mode", Claude Pro adds Research, which the Free plan does not have, and Google AI Pro can generate Deep Research reports with the Pro model and higher report limits, according to Google's Deep Research help page. None of the three vendors publishes a monthly report count for these plans.
Coding is where the plans diverge most. Claude Pro includes Claude Code in the terminal, IDEs and the Claude apps, drawing on the same allowance as chat. ChatGPT Plus includes Codex on the web, in the CLI, in the IDE extension and on iOS, with automatic code review and a Slack integration. Google AI Pro bundles entry rate limits for the agent model in Google Antigravity and premium Google Developer Program benefits, which include Gemini Code Assist with higher quotas and $10 of Google Cloud credits a month. Google's Jules coding agent, still listed in its AI Pro benefits, becomes read-only on November 11 and shuts down on December 31, as covered in our report on the Jules shutdown.
........
Capability | ChatGPT Plus | Claude Pro | Google AI Pro |
Research reports | Deep research, expanded | Research included | Deep Research with Pro model |
Coding agent | Codex (web, CLI, IDE, iOS) | Claude Code (terminal, IDE, apps) | Antigravity entry limits, Gemini Code Assist |
Browser agent | Agent mode; Chrome extension marked limited | Claude in Chrome | Chrome auto browse |
Office apps | Not listed for Plus | Microsoft 365 integration | Gemini in Gmail, Docs, Vids |
Memory and projects | Expanded memory, projects, custom GPTs | Memory, unlimited projects | Gems, Gemini Notebook |
Scheduled work | Tasks | Task handoff and scheduling | Scheduled actions |
........
Google's help pages set an age requirement of 18 for Deep Research, Antigravity, Jules, Colab and Tasks in Gemini on AI Pro, which matters for family members who share the plan.
··········
IMAGE, VIDEO AND CREATIVE TOOLS ON THE THREE PLANS.
Google AI Pro is the only plan of the three with video generation in the main app. Its subscriptions page lists video generation and "a limited trial of Veo 3.1 Lite", 1,000 Google Flow credits a month, and image generation with Nano Banana models. OpenAI's plan page lists "more complex and accurate image creation" for Plus and image generation with Thinking, and its comparison table shows voice with video on all paid plans; no video generation row appears for Plus. Anthropic's pricing page lists Artifacts, Claude Design, Docs and Slides, which are available from the Free plan, and no image or video generation feature for Pro.
A subscriber who wants to produce images and short videos with the same plan used for writing has a clear answer here, and it is Google AI Pro, with ChatGPT Plus as the alternative for still images only.
··········
STORAGE, FAMILY SHARING AND BUNDLED GOOGLE SERVICES.
Google AI Pro is a Google One plan, and that changes the value calculation. The plan includes 5 TB of storage across Google Drive, Gmail and Google Photos, family sharing of storage and AI features with up to five other people, YouTube Premium Lite, Google Home Premium Standard and Google Health Premium, according to Google One's help page on AI Pro benefits. Some benefits cannot be shared with the family group: the Google Developer Program premium benefit and YouTube Premium Lite stay with the plan manager.
Neither ChatGPT Plus nor Claude Pro includes storage beyond the files kept in chats and projects, and neither offers family sharing on its individual plan. Their extras are developer tools: Codex on Plus and Claude Code on Pro, against Gemini Code Assist and $10 of monthly Google Cloud credits on Google AI Pro.
A household already paying for Google One storage at the 2 TB level or above sees the largest effective discount, because AI Pro replaces that storage bill while adding the AI features.
··········
API PRICES OF THE MODELS BEHIND EACH PLAN.
The vendors' API price lists give a reference for what the flagship models cost outside a subscription. The prices below are standard paid-tier rates in USD per 1M tokens, taken from the OpenAI, Anthropic and Google developer pricing pages on November 4, 2026.
........
Model | Input | Cached input | Output | Notes |
GPT-6.1 Sol | $2.00 | $0.10 | $10.00 | 1,050,000-token context, 128K max output |
Claude Opus 5.5 | $4.00 | $0.20 | $20.00 | Batch $2 input, $10 output |
Claude Sonnet 5.5 | $2.00 | $0.10 | $10.00 | Batch $1 input, $5 output |
Gemini 4 Argon, through Nov 30 | $2.00 | $0.10 | $10.00 | Batch $1 input, $5 output |
Gemini 4 Argon, from Dec 1 | $4.00 | $0.20 | $20.00 | Batch $2 input, $10 output |
........
Google's pricing page shows Gemini 4 Argon at an introductory price that doubles on December 1, 2026, which brings it to the same list price as Claude Opus 5.5. GPT-6.1 Sol and Claude Sonnet 5.5 share identical list prices.
··········
WHAT A MONTH OF HEAVY USE WOULD COST ON THE API INSTEAD.
To show what the $20 buys, the calculation below prices two typical monthly workloads at the API rates above, in Python. Workload A is daily writing and analysis: 500 messages a month (20 on each of 25 working days), with 6,000 input tokens per message including conversation history and attached text, 900 output tokens, and half of the input served from cache. Workload B is a coding agent: 17 sessions a month (four a week), each with 1.5 million input tokens across its tool calls, 85 percent of them cache hits, and 60,000 output tokens. The calculation assumes every individual request stays within the standard price tier of each model.
def cost(input_price, cached_price, output_price, tokens_in, cached_share, tokens_out):
fresh = tokens_in * (1 - cached_share) * input_price
cached = tokens_in * cached_share * cached_price
return (fresh + cached + tokens_out * output_price) / 1_000_000
workload_a = (500 * 6_000, 0.50, 500 * 900) # 3.0M in, 0.45M out
workload_b = (17 * 1_500_000, 0.85, 17 * 60_000) # 25.5M in, 1.02M out........
Model at API rates | Workload A (chat) | Workload B (coding agent) | A + B |
GPT-6.1 Sol | $7.65 | $20.02 | $27.67 |
Claude Sonnet 5.5 | $7.65 | $20.02 | $27.67 |
Claude Opus 5.5 | $15.30 | $40.03 | $55.33 |
Gemini 4 Argon, through Nov 30 | $7.65 | $20.02 | $27.67 |
Gemini 4 Argon, from Dec 1 | $15.30 | $40.03 | $55.33 |
........
At these assumptions, a user who does both kinds of work every month would spend between $27.67 and $55.33 on the API, so each $20 plan is cheaper as long as its limits are not reached. The comparison shifts with the plan's allowance: nobody outside the vendors knows how many of Workload B's sessions fit into Claude Pro's weekly limit or Google AI Pro's four-times-standard allowance, while OpenAI's Plus estimates of 15 to 160 GPT-6.1 Sol messages per five hours allow a rough check against one's own usage.
··········
WHAT OPENAI, ANTHROPIC AND GOOGLE DO NOT PUBLISH ABOUT THESE PLANS.
The gaps in the official documentation are large enough to affect the choice. OpenAI's plan page for Plus still lists GPT-5.5 model labels while its Codex documentation lists the GPT-6 family, and it gives no figure for chat message limits or file upload sizes on Plus. OpenAI also announced on November 2 that the Chat and Work tabs will merge, covered in our article on the ChatGPT tab merge, so the current split between chat limits and the Work and Codex allowance is temporary.
Anthropic publishes no message count, token budget or weekly figure for Claude Pro, only the multiple of at least five times the Free plan per session, and it does not state which Opus, Sonnet and Haiku versions each plan uses at a given moment beyond its model pages. Google publishes multipliers instead of prompt counts, does not state a Gemini 4 Argon quota for AI Pro, and does not state the free-trial terms on its plan pages. Our guides to Claude plans and usage limits and to Gemini plans from Free to Ultra follow these limits plan by plan.
··········
CHATGPT PLUS VS CLAUDE PRO VS GOOGLE AI PRO: WHICH PLAN FITS WHICH USER.
The choice depends less on the headline price, which is the same, than on the kind of work and on what else the subscriber already pays for.
........
User or workload | Recommended plan | Reason |
Long documents and large files in chat | Claude Pro or Google AI Pro | 1M-token context against 256K on Plus |
Coding agent in the terminal with the top model | Claude Pro | Opus 5.5 and Claude Code inside the plan allowance |
Coding agent with published per-model limits | ChatGPT Plus | Codex estimates of 15-160 GPT-6.1 Sol messages per 5 hours |
Household or family use | Google AI Pro | Sharing with up to 5 people and 5 TB storage |
Gmail, Docs and Drive users | Google AI Pro | Gemini inside Google apps |
Microsoft 365 users | Claude Pro | Microsoft 365 integration listed for Pro |
Images and short video | Google AI Pro | Video generation, Flow credits, Nano Banana |
Lowest cost over a year | Claude Pro or Google AI Pro on annual billing | $200 and $199.99 against $240 for Plus |
........
Google AI Pro is the best value for most non-developers, because the storage and family sharing are worth close to the subscription price for anyone who would otherwise pay for cloud storage. Claude Pro is the stronger plan for long-context work and terminal coding with Opus 5.5, and its annual price matches Google's. ChatGPT Plus is the right choice for users who want the GPT-6 family in Codex with limits they can estimate in advance, provided they accept the 256K-token chat context and a monthly price with no annual discount.
··········
FOLLOW US FOR MORE.
·····
DATA STUDIOS
·····
datastudios.org
