Claude Max vs Ultra: is $200 worth it now? (And every other $100 vs $200 coding plan)
Short answer: on Oct 30, 2026 Anthropic turned Max 20x into Ultra ($200). Ultra gives you 2x Max's weekly usage (10x Pro vs 5x Pro) and, once your account is verified, no five-hour limit. So it's worth it if you hit Max's weekly cap or keep hitting the five-hour wall in the middle of long Claude Code runs. If you only hit the wall now and then, Max at $100 plus usage credits is cheaper. And across other tools, doubling from $100 to $200 buys anywhere from 2x to 4x more usage, depending on who you pay.
Checked Nov 4, 2026 on each vendor's own pages: Claude pricing, Claude Ultra help article, Claude API credits docs, ChatGPT Pro tiers, Codex pricing, Copilot plans, Cursor (X Subscription) pricing, Google AI plans, Antigravity plans, Kiro billing and Devin pricing. Our own math is labeled as an estimate.
Claude Max vs Ultra, side by side
| Max | Ultra | |
|---|---|---|
| Price | $100/mo, monthly only | $200/mo, monthly only |
| Five-hour limit | Yes, 5x Pro per five hours | None for verified accounts. New accounts keep a temporary one until Anthropic verifies them |
| Weekly limit | 5x Pro | 10x Pro (2x Max) |
| Fable (top model) | Up to half your weekly limit | Up to half your weekly limit |
| Monthly API credits | $100 | $200 |
| Claude Code | Included, shares the same limits | Included, shares the same limits |
| Past your limit | Usage credits at API rates (optional) | Usage credits at API rates (optional) |
| Old name | Max 5x | Max 20x (moved over automatically, same price) |
What actually changed for Max 20x people
- The headline number went from "20x" to "10x". That's not a cut on its face: the old 20x was per five-hour session, the new 10x is per week. Anthropic doesn't publish either allowance as a number, so nobody outside Anthropic can say for sure whether a week of work fits in more or less than before. Watch Settings > Usage for a couple of weeks before you decide.
- The five-hour wall is gone once you're verified. Check Settings > Usage: if there's no "Current session" bar, you're verified. Anthropic says verification is automatic, based on account and payment history, and verifying your phone at checkout doesn't count.
- The weekly cap is now the only cap. That cuts both ways: you can burn a whole day on one huge refactor, but you can also burn half your week in that one day. There's also a separate weekly limit for Fable.
Sources: claude.com/pricing and What is the Ultra plan?, both checked Nov 4, 2026. Anthropic also says it may apply other caps "at our discretion" on every plan.
Should you pay $200 for Ultra?
A quick way to decide, based on which limit you actually hit on Max. This is our reading of Anthropic's plan rules, not Anthropic guidance.
| What happens to you on Max | Our pick | Why |
|---|---|---|
| You rarely hit any limit | Stay on Max, or try Pro ($20) | You're paying for headroom you don't use |
| You hit the five-hour limit once or twice a week | Max + usage credits | A few hours at API rates usually costs less than another $100. Set a monthly spend limit |
| You hit the five-hour limit most days, but never the weekly one | Ultra | This is exactly what Ultra fixes |
| You hit the weekly limit before the week ends | Ultra | 2x the weekly allowance for 2x the price, plus no five-hour wall |
| You hit Ultra's weekly limit too | Ultra + usage credits, or split work to a second tool | There's no higher individual Claude plan. See the other $200 plans below |
Usage credits bill at standard API rates. For scale, our medium agent task estimate (1.5M input tokens, 90% cached, 25K output) is about $0.69 on Sonnet 5.5 and $1.37 on Opus 5.5. That's an estimate, and long Claude Code runs can cost far more per task (see cost per task).
The API credit catch: Max's $100 and Ultra's $200 don't work in Claude Code
This is the part people get wrong. Max and Ultra now come with $100 or $200 a month of Claude Platform credits, but Anthropic's docs say they don't apply to interactive Claude Code (terminal, IDE, desktop or web) and can't pay for usage credits when you run out in Claude or Claude Code.
- What they do cover: the Claude API, Message Batches, the Console Playground, Claude Managed Agents and the Claude Agent SDK.
- How to claim: after 7 days on the plan, go to Settings > Billing on claude.ai and link a Claude Console organization. No card needed on the Console side.
- Use it or lose it: credits refresh each billing cycle and expire at the end of it. No rollover.
- Not on Bedrock, Vertex AI or Foundry. Only Anthropic's own platform.
A way to get value from them (our idea, an estimate): put the credits to work on jobs that don't need interactive Claude Code, like a nightly Agent SDK script that triages issues, a batch job that writes tests, or a PR summary bot. At our medium-task estimate, $100 is roughly 145 Sonnet 5.5 tasks a month, and $200 roughly 290. We couldn't confirm whether running claude -p with an API key counts as "interactive Claude Code" for this rule, so stick to the Agent SDK or API if you want to be sure.
OpenAI is different here. ChatGPT Pro also includes $100, $200 or $500 a month in API credits. They don't add to signed-in Codex usage either, but Codex can run with an API key at API prices, so they can act as overflow for Codex CLI work (details on our ChatGPT Pro Codex limits page).
Sources: Monthly API credits for Max, Ultra, and Team plans and Claude Platform docs, checked Nov 4, 2026.
What going from $100 to $200 actually buys, tool by tool
Every big AI coding vendor now sells a $100 tier and a $200 tier. They're not equal. Here's how much more usage the second $100 gets you, using each vendor's own multipliers.
| Tool | ~$100 plan | ~$200 plan | Usage you get for 2x the price | Five-hour limit? | Claude Opus included? |
|---|---|---|---|---|---|
| Claude (Claude Code) | Max $100: 5x Pro | Ultra $200: 10x Pro weekly | 2x weekly, plus no five-hour limit | Max yes, Ultra no (verified) | Yes |
| OpenAI Codex (ChatGPT Pro) | Pro 100 | Pro 200 | More, but OpenAI doesn't say how much | No five-hour limit on Pro right now; weekly limits may apply | No (OpenAI models only) |
| Cursor (X Subscription) | Super $100: 15x Plus | Ultra $200: 50x Plus | 3.3x | No, one weekly pool | Not in the pool. Claude bills at API price from credits |
| Google Antigravity (Google AI Ultra) | Ultra $99.99: 5x Pro | Ultra $199.99: 20x Pro | 4x | Yes, refreshes every five hours up to a weekly limit | Yes, Ultra includes third-party models |
| Kiro | Pro Max $100: 5,000 credits | Power $200: 10,000 credits | 2x (every plan is $0.02 a credit) | No, monthly credits | Claude models on paid plans (see Kiro's models page) |
| GitHub Copilot | Max $100: $200 in AI credits | No $200 plan. Max + $100 extra usage = $300 in credits | 1.5x | No, monthly credits | Yes (Pro+ and up) |
| Devin | No $100 plan | Max $200 | "Significantly higher quotas" than Pro $20, no number | Daily and weekly allowance | Frontier Claude models on paid plans |
Multipliers are each vendor's own: Claude (pricing), Cursor (X Subscription pricing: Plus base, Premium 4x, Super 15x, Ultra 50x), Google (Google AI plans: 5x or 20x Pro), Kiro (billing docs: Pro 1,000, Pro+ 2,000, Pro Max 5,000, Power 10,000 credits), Copilot (plans: Max is $100 base plus a $100 "flex allotment" GitHub says may change), OpenAI (Pro tiers, Codex pricing), Devin (pricing). The multipliers compare tiers within one tool only. A "15x" in Cursor and a "5x" in Claude are not the same amount of work.
What jumps out
- Best step-up: Google AI Ultra (4x for 2x) and Cursor (3.3x for 2x). If you're already on their $100 tier and running out, the $200 tier is the better deal per dollar. With Cursor, remember Claude and GPT models aren't in that pool at all.
- Claude is a straight 2x on the weekly cap. What you're really paying the extra $100 for is losing the five-hour limit.
- Only two of these drop the five-hour window: Claude Ultra (verified accounts) and ChatGPT Pro (all three tiers, for now). Google AI Ultra still refreshes every five hours.
- Copilot Max is the only $100 plan that tells you its allowance in dollars: $200 of AI credits a month. At our medium-task estimate that's about 290 Sonnet 5.5 tasks or 145 Opus 5.5 tasks (an estimate, assuming credits track API list prices). Half of it is the flex allotment, which GitHub says may change.
- Kiro has no bulk discount. Every plan works out to $0.02 a credit, half the $0.04 add-on price. Pick the tier that matches your month; bigger tiers aren't cheaper per credit.
- ChatGPT Pro is the only one with a $500 tier. Pro 500 adds Ultrafast mode, which uses your allowance at 8x the normal rate.
Who should pick what at $100 to $200
- You live in Claude Code and the five-hour wall is the problem: Claude Ultra ($200). Nothing else gives you Claude Code with no five-hour window.
- You want Claude Opus, but $200 feels steep: Copilot Max ($100, $200 in credits, Opus included) or Google AI Ultra $99.99 (Opus 5.5 in Antigravity). You lose Claude Code itself; see Claude Code alternatives.
- You use OpenAI models and long sessions: ChatGPT Pro 100. No five-hour Codex limit right now, and its allowance doubled on Oct 29. Try it before Pro 200.
- You use Cursor's own models (Composer, Grok) all day: Cursor Ultra ($200) gives the biggest step-up of the bunch at 50x Plus.
- You want a hard monthly number: Kiro or Copilot. Both publish credits, so you can plan your month.
- You're on the fence between two tools: splitting about $110 across two tools can beat $200 on one. For example, Claude Max ($100) for Claude Code plus Copilot Pro ($10) for completions and quick cloud tasks.
Our opinion based on the pages linked above, not vendor guidance.
What we couldn't confirm
- Whether Ultra's 10x weekly allowance is more or less work than the old Max 20x gave you over a week. Anthropic doesn't publish allowances as numbers.
- How long Ultra verification takes for a new account.
- Whether
claude -pwith an API key can use the Max and Ultra API credits. The docs exclude "interactive" Claude Code and list the Agent SDK as covered. - How much bigger ChatGPT Pro 200 is than Pro 100, and the size of Devin Max's quota. Neither vendor says.
- How long Copilot keeps the $100 flex allotment on Max.
Ready to try one?
Every tool here can be cancelled monthly, so the fastest way to decide is a real task on each.
Claude Code
OpenAI Codex
GitHub Copilot
Google Antigravity
FAQ
What's the difference between Claude Max and Ultra?
Max is $100 a month with 5x Pro usage every five hours and every week, plus $100 a month in API credits. Ultra is $200 a month with 10x Pro usage per week, no five-hour limit for verified accounts, and $200 a month in API credits. Both include Claude Code and Fable up to half the weekly limit, and both are monthly only.
Is Claude Ultra the same as Max 20x?
It replaced it. On Oct 30, 2026 Anthropic renamed Max 20x to Ultra at the same $200 price and moved subscribers over automatically. The limit is now described as 10x Pro per week with no five-hour limit, instead of 20x Pro per five-hour session. Anthropic doesn't publish either allowance as a number.
Does Claude Ultra really have no usage limit?
No. Verified Ultra accounts have no five-hour limit, but there's still a weekly limit across all models and a separate weekly limit for Fable, and Anthropic says it may add other caps. New accounts keep a temporary five-hour limit until Anthropic verifies them automatically.
Can I use the Max or Ultra API credits in Claude Code?
No. Anthropic says the monthly API credits don't apply to interactive Claude Code in the terminal, IDE, desktop or web, and can't pay for usage credits in Claude or Claude Code. They work on the Claude API, Message Batches, the Console Playground, Managed Agents and the Claude Agent SDK, and expire each billing cycle.
Is Ultra worth $200?
It's worth it if you hit Max's weekly limit, or hit the five-hour limit most days. If you only hit the five-hour wall once or twice a week, Max at $100 plus usage credits at API rates is usually cheaper. That's our reading of Anthropic's plan rules, not Anthropic guidance.
Which $200 AI coding plan gives the most extra usage over its $100 tier?
By each vendor's own multipliers: Google AI Ultra goes from 5x to 20x Pro (4x for double the price), Cursor from Super 15x to Ultra 50x Plus (3.3x), and Claude and Kiro double (2x). Copilot has no $200 plan, and OpenAI and Devin don't publish multipliers. These multipliers only compare tiers within one tool.
Ship your app: the rest of the stack
Picked your AI coding tool? These are the usual next costs. Free tiers first, and prices checked Oct 20, 2026 on each vendor's pricing page.
Code review
- CodeRabbitAI review on every pull request, in GitHub or GitLabFree for public repos; Essentials $24/user/mo billed annually ($30 monthly)
Hosting
- VercelNext.js and front-end appsHobby free (personal, non-commercial); Pro $20/mo per developer seat
- Netlify referralStatic sites and front ends, unlimited Pro membersFree; Personal $9/mo; Pro $20/mo with unlimited members
- RailwayBack ends, workers and databases with no server setupFree trial $5 credit; Hobby $5/mo incl. $5 usage; Pro $20/mo
- RenderHeroku-style web services, cron jobs and PostgresHobby $0 + compute; Pro $25/mo + compute
Database and auth
Domain name
- Cloudflare RegistrarDomains at cost with no markupWholesale price per TLD
- PorkbunCheap domains with free WHOIS privacyVaries by TLD
- NamecheapPopular registrar with hosting and email add-onsVaries by TLD
Editorial disclosure
Rankings are not for sale. Sponsored placements and affiliate links are always labeled, and they never change our recommendations, quiz results, prices, or comparison data. Links marked "referral" are affiliate links (Netlify): if you buy through them, we may earn a commission at no extra cost to you. Every other link is a plain link that pays us nothing, and pricing source links are never affiliate links. Editorial policy.