The AI coding subscription market reshuffled again in the last few weeks. Anthropic now sells Free, Pro, Max, and Ultra tiers with monthly API credits on the top two. GitHub added a $100 Copilot Max plan. Cursor moved to a five-tier structure sold as part of The X Subscription and renamed its old plans legacy. Trae replaced its $3 and $10 plans with $20, $60, and $200 tiers that publish a dollar amount of usage. This guide lines up the plans from the official pricing pages as of November 13, 2026, explains how to compare them when every vendor measures usage differently, and ends with a decision guide. Our subscription ROI tables carry the arithmetic.
Plans by price
Up to $10 per month
| Plan | Price | What the vendor says you get |
|---|---|---|
| GitHub Copilot Free | $0 | 2,000 completions per month, access to smaller models such as Haiku 4.5 and GPT-5 mini, Copilot CLI |
| Cursor Free | $0 | Limited Agent and Chat requests |
| Kiro Free | $0 | 50 credits per month, open-weight models and Claude Sonnet 4.5 |
| Google AI Plus | $4.99 | Entry Google AI plan |
| Cursor Plus | $8 | Base usage pool; Grok and Cursor models only, no third-party models |
| ChatGPT Go | $8 | More messages and uploads than Free; may include ads |
| GitHub Copilot Pro | $10 | Unlimited completions, $15 of monthly credits, third-party agents (Claude Code and Codex), cloud agent and code review |
| OpenCode Go | $10 | Open coding models with dollar-based allowances |
$17 to $20 per month
| Plan | Price | What the vendor says you get |
|---|---|---|
| Claude Pro | $20 monthly, or $17 per month billed annually ($200 up front) | More usage than Free, Claude Code included, Fable paid with usage credits |
| ChatGPT Plus | $20 | Expanded messages, deep research, Codex and ChatGPT Work usage |
| Google AI Pro | $19.99 | Expanded Gemini access, 5 TB storage, $10 in monthly Google Cloud credits |
| Trae Pro | $20 | All models, $20 of usage per month, up to 10 concurrent cloud tasks |
| Windsurf Pro | $20 | Individual plan on the Windsurf pricing page |
| Kiro Pro | $20 | 1,000 credits per month, add-on credits at $0.04 each |
$30 to $60 per month
| Plan | Price | What the vendor says you get |
|---|---|---|
| Cursor Premium | $30 | 4x Plus usage; cloud agents, automations, code hosting, third-party models at API list price from prepaid credit |
| GitHub Copilot Pro+ | $39 | Premium models including Opus, 4x+ the usage of Pro, $70 of monthly credits |
| Kiro Pro+ | $40 | 2,000 credits per month |
| OpenCode Go Plus | $40 | Higher limits on the same open models as Go |
| Trae Pro+ | $60 | $60 of usage per month, up to 15 concurrent cloud tasks |
$100 per month
| Plan | Price | What the vendor says you get |
|---|---|---|
| Claude Max | $100 | 5x Pro usage, Fable included up to half of the weekly limit, $100 per month in API credits |
| ChatGPT Pro (entry tier) | From $100 | Choice of three usage tiers, Pro reasoning, longer Codex and ChatGPT Work sessions, up to $500 of monthly API credits |
| GitHub Copilot Max | $100 | 2.9x+ the usage of Pro+, priority access to new models, $200 of monthly credits |
| Cursor Super | $100 | 15x Plus usage, priority model access at peak times |
| Google AI Ultra | From $99.99 | 5x higher Gemini limits than AI Pro, larger Colab and Cloud credits |
| Kiro Pro Max | $100 | 5,000 credits per month |
$200 per month
| Plan | Price | What the vendor says you get |
|---|---|---|
| Claude Ultra | $200 | No 5-hour limit for verified accounts, 10x Pro weekly usage, $200 per month in API credits |
| Cursor Ultra | $200 | 50x Plus usage, earliest access to new products and models |
| Trae Ultra | $200 | $200 of usage per month (10x Pro), model early access, up to 20 concurrent cloud tasks |
| Windsurf Max | $200 | Top individual plan on the Windsurf pricing page |
Our catalog tracks ChatGPT Pro at three price points ($100, $200, and $500) and notes that the $200 tier was reduced to 10x Plus usage on October 30, 2026. Check OpenAI's pricing page for the current tier names before you buy.
What changed this fall
- Anthropic: the former Max 5x and Max 20x structure is now Max ($100, 5x Pro) and Ultra ($200, no 5-hour limit for verified accounts, 10x Pro weekly). Both include monthly API credits for the Claude Platform ($100 and $200). The Fable model is included on Max and Ultra up to half of the weekly limit, and paid with usage credits on Pro.
- GitHub: Copilot now has three paid plans, each with a monthly credit amount: Pro $15, Pro+ $70, Max $200.
- Cursor: sold as part of The X Subscription, one subscription for Grok, Cursor, and X. Paid plans share a weekly pool of usage for Cursor models, and third-party models are charged at API list price from prepaid credit on Premium and above. The former Pro, Pro+, and Ultra plans are called legacy and keep working for existing subscribers.
- Trae: the Lite plan is gone and the tiers publish dollar usage: $20, $60, and $200.
- Google Antigravity: Gemini CLI stopped serving free and AI Pro users on June 18, 2026, and Antigravity CLI replaced it. The Antigravity free plan has basic weekly limits and Google AI Pro and Ultra add more.
How to compare plans fairly
Vendors measure usage in incompatible units: messages, requests, credits, dollars, and multiples of a base plan. Three habits make comparisons honest.
- Convert to dollars where the vendor publishes them. Trae publishes $20, $60, and $200 of usage. GitHub publishes monthly credits of $15, $70, and $200. Anthropic and OpenAI list monthly API credits on their top tiers. Dollar amounts compare directly with the price you pay.
- Treat multiples as relative, not absolute. Cursor's 4x, 15x, and 50x are multiples of Plus, whose dollar size is not published. They tell you how the tiers relate to each other, not how much work each can do. For that reason our ROI tables keep Cursor's legacy plans and do not yet score the new tiers.
- Account for what is not in the pool. Cursor Plus is limited to Grok and Cursor models. Zed-style credits, Copilot credits, and Kiro credits are all consumed at different rates by different models.
Our methodology page explains how we estimate an API-equivalent value for a plan, and every plan page states its assumptions and the verification date.
Which plan for which developer
| You are | Consider |
|---|---|
| Trying AI coding for the first time | Copilot Free, Kiro Free, Cursor Free, or free models on OpenCode Zen |
| A student | The student offers in our student guide before you pay |
| A hobbyist or occasional coder | Copilot Pro ($10), OpenCode Go ($10), or Claude Pro or ChatGPT Plus ($20) |
| A daily agent user | Claude Max, ChatGPT Pro, Copilot Pro+ or Max, Cursor Super |
| Running agents for most of the workday | Claude Ultra, ChatGPT Pro higher tiers, Cursor Ultra |
| Budget-driven with open models | OpenCode Go, Z.ai Coding Plan (from $18), or metered API: see DeepSeek V4.1 Flash and GLM 5.3 |
Cost traps to watch
- Annual versus monthly. Claude Pro is $17 per month billed annually and $20 monthly. Cursor's annual prices are $80, $300, $1,000, and $2,000 for Plus, Premium, Super, and Ultra.
- Pay-as-you-go on top. Cursor charges third-party models at API list price from prepaid credit, so the plan price is not the whole bill.
- Auto-renewing promotions. Free-year offers renew at full price unless you cancel.
- Plans change in months, not years. Check the date on any comparison, including this one.
FAQ
What is the cheapest paid AI coding plan?
Among the plans we track, Cursor Plus and ChatGPT Go are $8, and GitHub Copilot Pro and OpenCode Go are $10. They buy very different things: Plus is limited to Grok and Cursor models, Copilot Pro includes $15 of monthly credits, and Go covers open models.
What happened to Claude Max 20x?
Anthropic's pricing page now lists Max at $100 and Ultra at $200. Ultra has no 5-hour limit for verified accounts and 10x Pro weekly usage.
Does GitHub Copilot have a $100 plan?
Yes. Copilot Max is $100 per user per month with $200 of monthly credits and 2.9x or more of the usage of Pro+.
Are Cursor's old Pro plans gone?
Cursor says plans such as Pro, Pro+, and Ultra are legacy plans that keep working for existing subscribers. New subscriptions use the Free, Plus, Premium, Super, and Ultra tiers.
Which plan has the best value?
It depends on how much you use it. Our ROI tables estimate the API-equivalent value of each plan, and each plan page lists the assumptions.
How often do you update this?
Plan pages show a verification date. This article reflects the vendors' pages on November 13, 2026.
