Claude Pro vs Max vs Ultra: Which Claude Plan to Buy (2026)
The short answer
Buy the cheapest plan whose limit you do not hit. As of November 1, 2026:
- Pro ($20/month, $17 annual) — daily chat, writing and light Claude Code use.
- Max ($100/month) — all-day work: 5x Pro usage per five-hour window and per week, Fable included up to half the weekly limit, $100 a month of API credits.
- Ultra ($200/month) — the longest agentic sessions: 10x Pro’s weekly usage and no five-hour limit for verified accounts, plus $200 a month of API credits.
Ultra is the renamed Max 20x. Anthropic changed the names on October 30, 2026; prices did not change, and Max 20x subscribers moved to Ultra automatically.
The comparison
| Pro | Max | Ultra | |
|---|---|---|---|
| Price | $20/mo ($17/mo billed annually) | $100/mo | $200/mo |
| 5-hour limit | 1x | 5x Pro | None (verified accounts) |
| Weekly limit | 1x | 5x Pro | 10x Pro |
| Fable (top model) | Paid with usage credits | Up to 50% of weekly limit | Up to 50% of weekly limit |
| Monthly API credits | — | $100 | $200 |
| Claude Code | ✅ | ✅ | ✅ |
| Billing | Monthly or annual | Monthly | Monthly |
All three plans include Claude Code, Projects, Claude in Chrome and Microsoft 365, research, memory and a context window up to 1M tokens depending on the model.
What the five-hour limit change means
Until October 30, every paid plan was metered twice: a rolling five-hour session limit and a weekly limit. Heavy Claude Code users hit the five-hour wall mid-task and waited. Ultra removes the five-hour limit for verified accounts, leaving only the weekly allowance (and a separate weekly cap for Fable). Anthropic verifies accounts automatically from account and payment history; until then a temporary limit applies. You can check under Settings > Usage — if there is no “Current session” bar, the account is verified. Verifying a phone number at checkout does not remove the limit.
For people who run long autonomous coding sessions, that is the real difference between Max and Ultra — more than the multiplier.
The API credits, read carefully
Max and Ultra now include monthly credits for the Claude API: $100 and $200. Team plans get $20 per Standard seat and $100 per Premium seat, pooled and capped at $500 a month. Details that matter:
- Credits cover the Claude API, Claude Managed Agents, the Claude Agent SDK and the Console playground — not Claude Code and not extra usage in the Claude apps.
- They work only in the Claude Console, not on AWS, Google Cloud or Microsoft Foundry.
- They refresh each billing cycle and expire at its end — no rollover.
- New subscribers can claim them after seven days; Free, Pro and Enterprise are not eligible.
At current list prices, $200 buys about 20 million Claude Sonnet 5.5 input tokens ($2/$10 per million) or 10 million Claude Opus 5.5 input tokens ($4/$20); see our current API prices. For a side project or an internal agent, that can be most of a month’s bill.
Claude Ultra vs ChatGPT Pro
OpenAI made the same move in October 2026: ChatGPT Pro now comes in $100, $200 and $500 tiers with $100, $200 and $500 of monthly OpenAI API credits, and Business seats earn $20 (Standard) or $100 (Premium) per seat, capped at $500 per workspace. The plans are now near mirror images at $100 and $200. Choose by the agent you use all day — Claude Code points to Max or Ultra, Codex to ChatGPT Pro — and only ChatGPT has a $500 tier. Our older ChatGPT Pro 500 vs Pro 200 vs Pro 100 page covers the OpenAI side.
How to choose
- You hit the five-hour limit on Pro once a week or less: stay on Pro and buy usage credits when needed.
- You hit it daily, but rarely run out for the week: Max.
- You run multi-hour Claude Code sessions or parallel agents: Ultra — the missing five-hour limit is the feature.
- You also build on the API: count the credits; on Ultra they offset $200 of real API spend.
Last verified: November 1, 2026.