Claude Max vs Ultra vs ultracode: two plans, one Claude Code setting (2026-11)
As of 2026-11-12, Max ($100 per month) and Ultra ($200 per month) are Claude subscription plans on the claude.com pricing page, billed monthly with officially stated usage limits; ultracode is a Claude Code setting, not a plan, and you turn it on with /effort ultracode. Below, the official wording is compared item by item on what each one is, where you set or buy it, and whether you spend a monthly fee or tokens, plus why the max effort level and the Max plan share a name but are different things.
Updated 2026-11-13
Switch models on the same endpoint and measure the cost per token yourself; check /models for callable models.
- Billed per token — live rates on /models
- Pay by card (Visa / Mastercard), Apple Pay, Google Pay or crypto
- Self-service: top up and activate instantly
Four official points
Max and Ultra monthly fees
The claude.com pricing page lists Max at $100 'Per month' and Ultra at $200 'Per month'. Both are subscription plans billed monthly; the page also notes that prices shown do not include applicable tax and that price and plans are subject to change at Anthropic's discretion.
Official usage multiples and limits
The pricing page states Max has '5x Pro usage every 5 hours and every week', while Ultra has '10x Pro usage every week' and 'No 5-hour limit‡', with the footnote 'No 5-hour limit for verified accounts.' The help center adds that Max and Ultra have a separate weekly limit for Fable and that Claude and Claude Code share the same limits.
Version needed for the ultracode toggle
The Claude Code docs state that /effort ultracode off, the slider toggle, and keeping ultracode on at levels other than xhigh require v2.1.284 or later; passing ultracode to the --effort flag requires v2.1.203 or later.
Whether you spend a monthly fee or tokens
A plan is billed monthly and subject to officially stated limits; Anthropic's help center states that when ANTHROPIC_API_KEY is set, Claude Code authenticates with that key instead of a subscription plan and incurs API usage charges billed at standard API rates distinct from plan pricing. On QCode, billing is per token, with per-model rates on /models.
What each one actually is
As of 2026-11-12, Max and Ultra are two subscription plans on the claude.com pricing page: Max is $100 per month and Ultra is $200 per month, billed monthly and subject to officially stated usage limits; ultracode is a setting in Claude Code that, when on, orchestrates dynamic workflows for substantive tasks, and is not a plan. The names overlap: the Max plan shares its name with the max effort level in Claude Code, but one is a subscription and the other a reasoning depth.
'ultracode vs max' has a second meaning
People searching 'ultracode vs max' often mean two different things: the relationship between the Max plan and ultracode, and the fact that Claude Code's effort levels also include one called max. The official docs describe low, medium, high, xhigh and max: low suits quick exchanges you review as you go; medium suits day-to-day engineering work with a clear scope, such as implementing a new feature; high suits work where verification and edge cases matter; xhigh is deeper reasoning at higher token spend; max is the deepest reasoning level, which Anthropic notes may show diminishing returns and overthink, so test it before adopting broadly. Ultracode is not a level but a toggle that makes Claude plan a dynamic workflow for each substantive task at any effort level. The max level and the Max plan only share a name.
Timeline
Claude Code 2.1.160 set the dynamic-workflow trigger keyword to ultracode; the official changelog notes that the word workflow no longer triggers a run, while describing a workflow in your own words still works.
Claude Code 2.1.284 made ultracode its own toggle in /effort; the official changelog says it no longer forces xhigh effort and stays on at any effort level.
Checked 2026-11-12: the claude.com pricing page currently lists Max at $100 per month and Ultra at $200 per month; the current help-center and Claude Code docs wording, plus QCode's /models and endpoints docs, were checked the same day.
Published officially vs not stated
Confirmed from official wording
① The pricing page lists Max at $100 per month and Ultra at $200 per month; Max has 5x Pro usage every 5 hours and every week, Ultra has 10x Pro usage every week and a note of no 5-hour limit, with the footnote limiting this to verified accounts; the two plans include $100 and $200 a month in Claude Platform API credits respectively. ② The help center states Pro and Max have a five-hour session limit and a weekly limit, Ultra has a weekly limit and verified accounts have no five-hour session limit, and Max and Ultra also have a separate weekly limit for Fable. ③ The Claude Code docs state ultracode is a setting rather than an effort level and orchestrates dynamic workflows for substantive tasks; it can be turned on via /effort, a launch flag or a settings file. ④ The official changelog says 2.1.160 set ultracode as the trigger keyword and made it its own toggle in 2.1.284.
No official verdict on which is stronger
Anthropic has not put ultracode and max through one shared evaluation, and has published no verdict on which is stronger. The official docs describe the levels separately: max is the deepest reasoning level, which may show diminishing returns and overthink; ultracode is a toggle that orchestrates dynamic workflows and can stay on at any effort level. The docs also state the effort scale is calibrated per model, so the same level name does not represent the same underlying value across models. These are official descriptions of the levels, not a ranking.
How to tell them apart and use them
Paying monthly for Claude and Claude Code
Max and Ultra are subscription plans on claude.com at $100 and $200 per month on the official pricing page; one subscription covers Claude on web, desktop and mobile plus Claude Code in the terminal, subject to the official 5-hour and weekly limits. The help center states these limits are shared across Claude and Claude Code, so activity in both counts against the same limits. For the Fable weekly limit on each plan, see /claude-fable-5-2-plan-limits.
Setting effort levels in Claude Code with an API key
With an API key, effort levels and the ultracode toggle are still set inside Claude Code, and you spend tokens rather than a monthly fee. The help center states that with ANTHROPIC_API_KEY set, Claude Code authenticates with that key instead of a subscription and incurs API usage charges; the official workflows doc says dynamic workflows are available on all paid plans, with Anthropic API access, and on the cloud platforms it lists. On QCode, billing is per token, with per-model rates on /models.
How to turn on ultracode and get started
① In a Claude Code session, run /effort ultracode to turn it on or /effort ultracode off to turn it off; or press Tab in the /effort slider to flip the Ultracode toggle, then Enter. ② Launch with claude --effort ultracode, or set "ultracode": true in a settings file; the flag starts that session at xhigh effort. ③ Version requirements: /effort ultracode off and the slider toggle need v2.1.284 or later, while passing ultracode to the --effort flag needs v2.1.203 or later. ④ The official docs say ultracode is unavailable when workflows are turned off or the model does not support xhigh effort. For per-model detail on effort levels and thinking, see /claude-code-ultracode-guide and /claude-adaptive-thinking-effort-guide; for how plans relate to API credits, see /claude-pro-max-vs-api-key.
Using an API key on QCode
On QCode, set the Base URL to https://api.qcode.cc/api in Claude Code and other Anthropic-protocol tools and set the model to a Claude model listed on /models (for example claude-opus-5-5); one QCode key then calls them all. The same key also calls GPT and Chinese models through the OpenAI-compatible endpoint https://api.qcode.cc/openai/v1. Effort levels and the ultracode toggle are set inside Claude Code; QCode bills per token, with per-model rates on /models. For one-time quota without a monthly plan, see /pay-as-you-go. Claude models use the Anthropic endpoint.
Frequently asked questions
Is ultracode a subscription plan?
No. Ultracode is a setting in Claude Code, not a subscription plan; Max and Ultra are the subscription plans on the claude.com pricing page. The official docs say ultracode is turned on or off in /effort and, when on, orchestrates dynamic workflows for substantive tasks.
ultracode vs max: which is stronger?
Anthropic has published no verdict on which is stronger. Ultracode is a toggle that makes Claude plan a dynamic workflow for each substantive task and can stay on at any effort level; max is the deepest reasoning level in the official docs, which notes it may show diminishing returns and overthink, so test it before adopting broadly. The docs also state the effort scale is calibrated per model, so the same level name does not represent the same underlying value across models. The two are different kinds of settings.
What is the difference between Max and Ultra?
On the claude.com pricing page, Max is $100 per month with 5x Pro usage every 5 hours and every week; Ultra is $200 per month with 10x Pro usage every week and a note of no 5-hour limit, with the footnote limiting this to verified accounts. The help center states Pro and Max have a five-hour session limit and a weekly limit, Ultra has a weekly limit and verified accounts have no five-hour session limit, and Max and Ultra also have a separate weekly limit for Fable. The pricing page also says Ultra includes $200 a month in API credits, up from $100 on Max. Per Anthropic's help center, the Max 20x plan is now called the Ultra plan, with no price change (as of 2026-11-12).
Do I need the Ultra plan to use ultracode?
No. Ultracode is a Claude Code setting, not a feature handed out by plan. The official docs say it orchestrates dynamic workflows for substantive tasks and can stay on at any effort level; the official workflows doc says dynamic workflows are available on all paid plans, with Anthropic API access, and on the cloud platforms it lists, and that on Pro you turn them on from the Dynamic workflows row in /config. With an API key, the help center states Claude Code authenticates with that key instead of a subscription and incurs API usage charges.
Does ultracode use more tokens?
Anthropic has not quantified ultracode's token spend on its own. The official docs say each effort level trades token spend against capability: xhigh is deeper reasoning at higher token spend, and max is the deepest reasoning level, which may show diminishing returns and overthink. With ultracode on, the official docs say Claude plans a dynamic workflow for each substantive task at any effort level. The docs also state the effort scale is calibrated per model.
Do effort levels still work with an API key?
Yes. Effort levels and the ultracode toggle are both set inside Claude Code, regardless of whether you authenticate with a subscription or an API key. The help center states that with ANTHROPIC_API_KEY set, Claude Code authenticates with that key instead of a subscription plan and incurs API usage charges, billed at standard API rates distinct from plan pricing. On QCode, billing is per token, with per-model rates on /models.
Sources
Anthropic official sources: the claude.com pricing page and help-center articles 'Use Claude Code with your Pro, Max, or Ultra plan', 'What is the Ultra plan?' and the monthly API credits article. Claude Code official sources: the model-config docs, the workflows docs, the settings reference and the changelog. QCode: qcode.cc /models and /pay-as-you-go, plus the endpoints page on docs.qcode.cc. All checked on 2026-11-12.
Buy a one-time quota and start anytime
Not ready for a monthly plan? Pay-as-you-go Packs start at $4 for a one-time quota you use within 7 days; if you already have a Personal Plan, add a booster pack when you need more quota.
Related
One-time quota packs: no monthly plan, valid 7 days
One-time quota packs valid for 7 days, no subscription; per-model rates are on /models.
The ultracode toggle and effort levels
The ultracode toggle itself and how each effort level works.
How subscription plans relate to API credits
Subscription plans, API credits and how API key billing differs.
Crypto payment guide
Nine coins, 56 networks, zero platform fee for AI coding API credits.
Official statements checked on 2026-11-12 against Anthropic and Claude Code pages; model availability and rates follow qcode.cc/models.