What AI coding
really costs
Every plan and token price for Claude Code, OpenAI Codex, Cursor, GitHub Copilot, Devin and Google Antigravity, copied from each official pricing page and checked by hand.
Key findings
- $20 is the standard price of a serious plan. Claude Pro, ChatGPT Plus (which includes Codex) and Devin Pro all cost $20 a month. OpenAI and Cursor now sell $8 entry plans.
- Heavy users pay $100 to $200. Anthropic, Cursor and Devin top out at $200 a month for one person. OpenAI goes further: ChatGPT Pro comes at $100, $200 or $500.
- Only GitHub tells you the allowance in dollars. Copilot Pro includes $15 of AI credits. Anthropic and Cursor describe limits as multiples ("5x Pro", "50x Plus") and OpenAI gives ranges, such as 15 to 160 messages every 5 hours. From the pricing pages alone, you cannot compare how much work a plan buys.
- The model you pick matters more than the plan. Output tokens cost from $0.50 per million (Claude Haiku 5.5 and GPT-6 Luna) to $50 (GPT-6 Astra), a 100x spread.
- Prices are moving. Google's introductory prices on Gemini 4 Carbon and Argon double on 1 December 2026, and Gemini 3.8 Flash doubles on 1 January 2027. Windsurf's pricing page now sends you to Devin.
Plans and limits
31 plans
| What you get | Extra usage | Source | |||||
|---|---|---|---|---|---|---|---|
| Claude Code | Free | $0 | Not listed | Individual | Claude Code is not included on Free. | Not listed | claude.com checked 2026-11-14 |
| Claude Code | Pro | $20 | $17 | Individual | More usage than Free, reset every 5 hours and every week. | Not listed | claude.com checked 2026-11-14 |
| Claude Code | Max | $100 | Not listed | Individual | 5x Pro usage every 5 hours and every week. Includes $100 a month in API credits. | Not listed | claude.com checked 2026-11-14 |
| Claude Code | Ultra | $200 | Not listed | Individual | 10x Pro usage every week, no 5-hour limit for verified accounts. Includes $200 a month in API credits. | Not listed | claude.com checked 2026-11-14 |
| Claude Code | Team, standard seat | $25 | $20 | Team | More usage than Pro. 2 to 150 users. | Not listed | claude.com checked 2026-11-14 |
| Claude Code | Team, premium seat | $125 | $100 | Team | 5x the usage of a standard seat. Up to $500 a month in pooled API credits. | Not listed | claude.com checked 2026-11-14 |
| Claude Code | Enterprise | Custom | $20 | Team | Seat fee plus all usage at API rates. 20-seat minimum. | Usage billed at API rates | claude.com checked 2026-11-14 |
| OpenAI Codex | Free | $0 | Not listed | Individual | Quick coding tasks on GPT-6 Luna. | Not listed | learn.chatgpt.com checked 2026-11-14 |
| OpenAI Codex | Go | $8 | Not listed | Individual | Lightweight coding tasks. | Not listed | learn.chatgpt.com checked 2026-11-14 |
| OpenAI Codex | Plus | $20 | Not listed | Individual | Estimated local messages per 5 hours: 15 to 160 on GPT-6.1 Sol, 35 to 285 on GPT-6.1 Terra, 5 to 45 on GPT-6 Astra. Weekly limits may also apply. | Buy ChatGPT credits | learn.chatgpt.com checked 2026-11-14 |
| OpenAI Codex | Pro | $100 | Not listed | Individual | No 5-hour limit. Pro also comes at $200 and $500 a month with more usage. | Buy ChatGPT credits | learn.chatgpt.com checked 2026-11-14 |
| OpenAI Codex | Business | $25 | $20 | Team | Per-seat limits similar to Plus, larger cloud machines. 2+ users. | Workspace credits | learn.chatgpt.com checked 2026-11-14 |
| OpenAI Codex | API key | Pay per token | Not listed | Individual | No subscription. Every token billed at API prices. | Pay as you go | learn.chatgpt.com checked 2026-11-14 |
| Cursor | Free | $0 | Not listed | Individual | Limited AI usage. | Not listed | cursor.com checked 2026-11-14 |
| Cursor | Plus | $8 | Not listed | Individual | Shared AI usage pool. | Top-offs, or wait for the weekly refresh | cursor.com checked 2026-11-14 |
| Cursor | Premium | $30 | Not listed | Individual | 4x the usage of Plus. | Top-offs, or wait for the weekly refresh | cursor.com checked 2026-11-14 |
| Cursor | Super | $100 | Not listed | Individual | 15x the usage of Plus. | Top-offs, or wait for the weekly refresh | cursor.com checked 2026-11-14 |
| Cursor | Ultra | $200 | Not listed | Individual | 50x the usage of Plus. | Top-offs, or wait for the weekly refresh | cursor.com checked 2026-11-14 |
| Cursor | Teams | $30 | Not listed | Team | From $30 per user. 4x to 50x Plus usage per user. | Not listed | cursor.com checked 2026-11-14 |
| GitHub Copilot | Free | $0 | Not listed | Individual | 2,000 completions a month, limited chat and agent mode. | Not listed | github.com checked 2026-11-14 |
| GitHub Copilot | Pro | $10 | Not listed | Individual | $15 of AI credits a month ($10 base plus $5 flex). | $0.01 per extra AI credit | github.com checked 2026-11-14 |
| GitHub Copilot | Pro+ | $39 | Not listed | Individual | $70 of AI credits a month ($39 base plus $31 flex). | $0.01 per extra AI credit | github.com checked 2026-11-14 |
| GitHub Copilot | Max | $100 | Not listed | Individual | $200 of AI credits a month ($100 base plus $100 flex). | $0.01 per extra AI credit | github.com checked 2026-11-14 |
| GitHub Copilot | Business | $19 | Not listed | Team | 1,900 AI credits per user a month, pooled across the organization. | $0.01 per extra AI credit | docs.github.com checked 2026-11-14 |
| GitHub Copilot | Enterprise | $39 | Not listed | Team | 3,900 AI credits per user a month, pooled. | $0.01 per extra AI credit | docs.github.com checked 2026-11-14 |
| Devin (incl. Windsurf) | Free | $0 | Not listed | Individual | Light quota. Unlimited inline edits and Tab completions. | Not listed | devin.ai checked 2026-11-14 |
| Devin (incl. Windsurf) | Pro | $20 | Not listed | Individual | Higher quotas with OpenAI, Claude and Gemini models, refreshed daily and weekly. | Extra usage at API prices | devin.ai checked 2026-11-14 |
| Devin (incl. Windsurf) | Max | $200 | Not listed | Individual | Significantly higher quotas. | Extra usage at API prices | devin.ai checked 2026-11-14 |
| Devin (incl. Windsurf) | Teams | $120 | Not listed | Team | $80 a month team fee plus $40 per full developer seat (shown here for one seat). Own quota per seat. | Extra usage at API prices | devin.ai checked 2026-11-14 |
| Google Antigravity | Individual | $0 | Not listed | Individual | Basic weekly rate limits, no numbers published. Unlimited Tab completions. | Not listed | antigravity.google checked 2026-11-14 |
| Google Antigravity | Business via Google Cloud | $30 | Not listed | Team | Standard and Plus plans from $30 per seat. A $0-seat pay-as-you-go plan also exists. | Consumption pricing on the pay-as-you-go plan | antigravity.google checked 2026-11-14 |
API token prices
What you pay when a tool runs on your own API key, or when a plan bills extra usage at API rates. A typical coding task reads far more than it writes, so input and cached input often decide the bill. Work out a task's cost.
| Notes | Source | |||||
|---|---|---|---|---|---|---|
| Anthropic | Claude Fable 5.2 | $8 | $0.20 | $40 | Same price at any prompt length. | platform.claude.com checked 2026-11-14 |
| Anthropic | Claude Opus 5.5 | $4 | $0.20 | $20 | Same price at any prompt length. | platform.claude.com checked 2026-11-14 |
| Anthropic | Claude Opus 5 | $5 | $0.50 | $25 | Same price at any prompt length. | platform.claude.com checked 2026-11-14 |
| Anthropic | Claude Sonnet 5.5 | $2 | $0.10 | $10 | Same price at any prompt length. | platform.claude.com checked 2026-11-14 |
| Anthropic | Claude Sonnet 4.6 | $3 | $0.30 | $15 | Same price at any prompt length. | platform.claude.com checked 2026-11-14 |
| Anthropic | Claude Haiku 5.5 | $0.10 | $0.01 | $0.50 | Prompts over 100K tokens: $0.50 in, $2.50 out. | platform.claude.com checked 2026-11-14 |
| Anthropic | Claude Haiku 4.5 | $1 | $0.10 | $5 | Same price at any prompt length. | platform.claude.com checked 2026-11-14 |
| OpenAI | GPT-6 Astra | $10 | $1 | $50 | Prompts over 272K tokens: $20 in, $75 out. | developers.openai.com checked 2026-11-14 |
| OpenAI | GPT-6.1 Sol | $2 | $0.10 | $10 | Prompts over 272K tokens: $4 in, $15 out. | developers.openai.com checked 2026-11-14 |
| OpenAI | GPT-6 Sol | $2 | $0.20 | $10 | Prompts over 272K tokens: $4 in, $15 out. | developers.openai.com checked 2026-11-14 |
| OpenAI | GPT-6.1 Terra | $1 | $0.05 | $5 | Prompts over 272K tokens: $2 in, $7.50 out. | developers.openai.com checked 2026-11-14 |
| OpenAI | GPT-6 Luna | $0.10 | $0.01 | $0.50 | Prompts over 272K tokens: $0.20 in, $0.75 out. | developers.openai.com checked 2026-11-14 |
| Gemini 4 Carbon | $2 | $0.10 | $10 | Intro price to 30 Nov 2026, then $4 in, $20 out. | aistudio.google.com checked 2026-11-14 | |
| Gemini 4 Argon | $2 | $0.10 | $10 | Intro price to 30 Nov 2026, then $4 in, $20 out. | aistudio.google.com checked 2026-11-14 | |
| Gemini 4 Hydrogen | $1.50 | $0.075 | $7.50 | Same price at any prompt length. | aistudio.google.com checked 2026-11-14 | |
| Gemini 3.1 Pro Preview | $2 | $0.20 | $12 | Prompts over 200K tokens: $4 in, $18 out. | aistudio.google.com checked 2026-11-14 | |
| Gemini 3.8 Flash | $0.75 | $0.075 | $3.75 | Price to 31 Dec 2026, then $1.50 in, $7.50 out. | aistudio.google.com checked 2026-11-14 |
Methodology
What is included
AI coding tools with a public price list: Claude Code (Anthropic), Codex (OpenAI), Cursor, GitHub Copilot, Devin (whose pricing page now also covers Windsurf) and Google Antigravity. We list every plan a person or team can buy without talking to sales, plus Enterprise rows where a price is published.
Where the numbers come from
Only the vendor's own pricing page or official documentation. No reseller pages, no press articles, no screenshots from social media. Each row links to its source and shows the day we checked it. If a vendor does not state a number, we write "Not listed" rather than estimate one.
How we report prices
US dollars per person per month, before tax, on monthly billing. The yearly column is the monthly equivalent when billed yearly, only where the vendor shows it. Usage limits are quoted the way the vendor states them: we do not convert "5x Pro" into a token count, because vendors do not publish one.
API token prices
Standard tier, per 1 million tokens, for current models. Where a vendor charges more for long prompts, the note shows the threshold and the higher price. Introductory prices are marked with their end date. Batch, priority and fast modes, regional surcharges and cache write prices are left out to keep the table comparable.
Updates
We recheck every source every week. When a price changes, the row and the date at the top change with it.
Cite or download
Jenzo Research (2026). AI coding price and limits index. https://jenzo.ai/research/ai-coding-prices (checked 14 November 2026).
Free to reuse with a link to this page.
Questions
How much does Claude Code cost?
Claude Code comes with Claude's paid plans: Pro at $20 a month ($17 a month billed yearly), Max at $100 and Ultra at $200. Team seats cost $25 a month ($20 yearly), or $125 for a premium seat with 5x the usage. You can also run it on API tokens, where Claude Sonnet 5.5 costs $2 per million input tokens and $10 per million output tokens.
How much does OpenAI Codex cost?
Codex is included in ChatGPT plans: Free, Go at $8, Plus at $20, Pro at $100, $200 or $500, and Business at $25 a seat ($20 billed yearly). With an API key you pay per token instead, from $0.10 per million input tokens on GPT-6 Luna to $10 on GPT-6 Astra.
Which AI coding tool is cheapest?
Every tool here has a free tier. The cheapest paid plans are Codex Go and Cursor Plus at $8 a month, then GitHub Copilot Pro at $10. The cheapest plan is not always the cheapest per task, because the usage you get differs a lot and is often not stated as a number.
What are the usage limits on Claude, Codex and Cursor?
Claude plans reset every 5 hours and every week, with Max at 5x Pro and Ultra at 10x Pro weekly. ChatGPT Plus gives an estimated 15 to 160 Codex messages every 5 hours on GPT-6.1 Sol, and Pro has no 5-hour limit. Cursor gives Premium 4x, Super 15x and Ultra 50x the usage of Plus, refreshed weekly.
How often is this index updated?
We check every official pricing page every week and update the date at the top. Each row links to its source and shows the day we checked it.