Inter-office memo
- To
- Everyone who wants Claude Opus for less
- From
- Mo, Chief Executive
- Date
- November 18, 2026, 1:47 AM CST
- Re
- The cheapest way to use Claude Opus 5.5
The cheapest way to use Claude Opus 5.5 for coding in 2026: 9 ways compared
For most people it’s Claude Pro, $20 a month or $200 a year, where Opus 5.5 is Claude Code’s default. Google AI Pro, $19.99, has it in Antigravity if you already pay Google; the API, $4 per million input tokens and $20 output, suits a few tasks a month; and Max, $100 for five times Pro’s usage, beats spending over $80 a month on usage credits. No free plan lets you choose Opus 5.5.
Disclosure: Enjoy is our product, and it runs Claude Code on your own Claude plan. We checked every plan and price below against Anthropic’s, Google’s, GitHub’s, Kiro’s and Cursor’s own pages on November 18, 2026, and we say where you don’t need Enjoy at all, which is most of the time.
Claude Opus 5.5 compared
Claude Opus 5.5 is what Anthropic calls its “Daily driver for agentic coding and enterprise work,” and Claude Code’s default: its model docs say “Pro, Max, Ultra, Team, Enterprise, and Anthropic API” all default to Opus 5.5. Anthropic isn’t the only place to get it. In short: Claude Pro is the cheapest way for most people, the API is cheaper only for a few tasks a month, and the other companies’ plans make sense when you already pay them. These are the ways to pay for it, side by side.
| Way | Best for | Price, checked November 18, 2026 | Where you use Opus 5.5 | How much you get |
|---|---|---|---|---|
| Claude Pro | Most people | $20 a month, or $200 a year | Claude Code and the Claude apps | Five-hour and weekly limits, shared with your chats |
| Claude Pro plus usage credits | Running out now and then | $20 a month, plus API rates past the limits | Claude Code and the Claude apps | Pro’s limits, then pay as you go |
| Claude Max or Ultra | Running out most weeks | $100 a month for Max, $200 for Ultra | Claude Code and the Claude apps | 5x Pro’s usage on Max; 10x Pro’s weekly usage on Ultra |
| The Anthropic API | A few tasks a month, and scripts | $4 per million input tokens, $20 per million output; no monthly fee | Claude Code with a Console account, and your own code | What you pay for |
| Google AI Pro | People who already pay Google | $19.99 a month, or $199.99 a year | Antigravity, on paid plans, not trials | A quota refreshed every five hours, until a weekly limit; no sizes published |
| GitHub Copilot Pro+ | People who already work in Copilot | $39 a month, with $70 of AI credits | Copilot Chat, its CLI, cloud agent and app | Credits, at $4 and $20 per million tokens |
| Kiro Pro | A free try for the rest of a month | $20 a month for 1,000 credits; the rest of your first month free | Kiro’s IDE, CLI and web app, and editors that support ACP | Credits, at twice the rate of Kiro’s Auto mode |
| Cursor Premium | People who already work in Cursor | $30 a month, plus Opus 5.5 at API list price | Cursor | Prepaid credit, at $4 and $20 per million tokens |
| Enjoy, with a Claude plan | Making a Claude plan go further | A monthly subscription on top of your plans, free for 7 days (pricing) | Claude Code, on your Claude plan | Your Claude plan’s |
Prices are each company’s US list prices, before tax. Anthropic bills Max and Ultra monthly only. Anthropic and Google don’t publish how many tokens their plans’ limits hold; GitHub, Kiro and Cursor bill Opus 5.5 in credits.
Is there a free way to use Opus 5.5?
Not one you can count on. Every free plan stops short of it:
- Claude Free has no Claude Code and no Opus: Anthropic’s pricing marks both “No” for Free.
- Antigravity’s free tier never had Opus 5.5, and its only Claude models, Sonnet 4.6 and Opus 4.6, are marked “Will be removed on November 18, 2026” on Google’s models page: today. A Google AI Pro trial doesn’t get Opus 5.5 either, which is “Available on Google AI Pro for non-trial subscriptions only.” Our memo has the details.
- GitHub Copilot Free can’t choose a model: “Copilot Student and Copilot Free users have access to models through auto model selection only,” say GitHub’s docs.
- Kiro Free includes “open weight models and Claude Sonnet 4.5, with limits,” says its pricing page, and its models page doesn’t offer Opus 5.5 on Free.
- Cursor’s Free and Plus plans “don’t include third-party models,” Claude among them.
The nearest thing is Kiro’s trial: “The first time you upgrade from Kiro Free to Pro, the rest of that month is free, and you get 500 trial credits to use within 12 months.” Pro includes Opus 5.5, and upgrading “requires a valid credit card.”
What Opus 5.5 costs, and why a plan usually wins
At Anthropic’s API prices, Opus 5.5 is $4 per million input tokens and $20 per million output, with cache reads at $0.20 per million and cache writes at $5. GitHub and Cursor charge the same $4 and $20. Coding agents read and write a lot: on Artificial Analysis’s comparison, one task in its Coding Agent Index takes Claude Code with Opus 5.5 at max effort 15.9 million tokens and costs $12.61 at API prices. Anthropic’s own costs page says that “across enterprise deployments, the average cost is around $13 per developer per active day and $150-250 per developer per month.”
Claude’s plans give much more than that for the money. SemiAnalysis, which measured how far each plan’s meter moves per token, found on October 5 that a $200 Claude plan “would still have 50% left after consuming $2,485 worth of Fable 5.1,” that Anthropic “offered the same per-dollar value for all of their subscription tiers,” and that “Opus 5.5 on any Claude subscription offers far better value than anything from OpenAI.” Anthropic publishes no token counts, and limits change, so treat that as one day’s measurement, not a promise.
So a Claude plan is the cheapest Opus 5.5 if you use it most days, the API if you use it a few times a month, and another company’s plan if you already pay it for something else.
Claude Pro
Best for: most people who want Opus 5.5 in Claude Code.
“Claude Code is included in all paid plans,” says Anthropic’s pricing page, Opus included, and on Pro its default model is Opus 5.5. It’s Anthropic’s cheapest plan with Opus 5.5, with the Claude apps on the same subscription.
Price: “$17 Per month with annual subscription discount ($200 billed up front). $20 if billed monthly.”
Watch for: the limits are shared and unpublished. Claude Code “shares the same usage limits as the rest of your plan, so your work in the terminal and your chats draw from one pool,” and Pro gets “More usage than Free, every 5 hours and every week,” with no number. Fable 5.2, Anthropic’s top model, is “paid with usage credits” on Pro.
Claude Pro plus usage credits
Best for: people who run out of Pro now and then.
Usage credits let Pro, Max and Ultra “continue using Claude after reaching their included usage limits,” says Anthropic’s help center, and they work in Claude Code too: “Usage credits apply to both Claude conversations and Claude Code terminal usage.” You set a monthly spend limit, and you can turn them off at any time.
Price: “Usage credits are billed at standard API rates,” so past your limits, Opus 5.5 costs what it does on the API.
Watch for: credits bought on or after December 7, 2026 “expire 12 months after the purchase date,” and “Usage credit purchases are non-refundable.” If they cost you more than $80 most months, Max is the better buy.
Claude Max or Ultra
Best for: people who run out of Pro most weeks.
Max is $100 a month for “5x Pro usage every 5 hours and every week.” Ultra is $200 for “10x Pro usage every week,” with “No 5-hour limit” for verified accounts. Both include Fable, “up to half your weekly limit,” and “Both plans are billed monthly.” Since SemiAnalysis measured the same value per dollar at every tier, Max isn’t a discount on Pro. It’s Pro’s rate with more room, and cheaper than buying the same work through usage credits at API rates.
Price: $100 or $200 a month.
Watch for: the API credits that come with them don’t pay for Claude Code. Max includes “$100 a month in API credits to build on the Claude Platform,” and Ultra $200, but Anthropic’s help center says “They don’t cover interactive Claude Code sessions or extra usage after you hit your plan’s limits.” They do cover claude -p run “with an API key from your linked Claude Console organization.” Our plan guide compares Pro, Max and Ultra.
The Anthropic API
Best for: a few Opus tasks a month, and scripts.
With a Claude Console account, Claude Code bills by the token, with no monthly fee: “For heavy coding sessions, you can also switch to pay-as-you-go API credits through a Console account,” says Anthropic’s pricing page. Prompt caching brings repeated context down to $0.20 per million tokens, and for work that can wait, the Batch API gives “a 50 percent discount on both input and output tokens,” or $2 and $10 per million for Opus 5.5.
Price: $4 per million input tokens and $20 per million output.
Watch for: it adds up fast in an agent. At Artificial Analysis’s $12.61 a task at max effort, two tasks cost more than a month of Pro. Fast mode for Opus 5.5 is “2x standard pricing.” And on a Claude plan, Anthropic’s help center warns that API credits in Claude Code are “billed at standard API rates (distinct from Pro/Max/Ultra plan pricing).”
Google AI Pro, in Antigravity
Best for: people who already pay for Google AI Pro, or want Gemini too.
Google’s models page lists Claude Opus 5.5 and Sonnet 5.5 in Antigravity on Google AI Pro, Ultra and Enterprise, on Google’s quota rather than Anthropic’s. Pro also brings Gemini 4 Argon and Gemini 4 Hydrogen, and the rest of Google AI Pro outside Antigravity.
Price: $19.99 a month, or $199.99 a year, on Google One. Ultra is from $99.99 a month, and “Ultra plans are billed monthly only.”
Watch for: Opus 5.5 is “Available on Google AI Pro for non-trial subscriptions only,” and Google publishes no quota sizes. Its plans page promises Pro “A high, generous quota, refreshed every five hours until the weekly limit is reached,” and says “Usage limits for this service are subject to modification.” Past the quota, AI credits “are consumed at standard Gemini Enterprise consumption pricing.” There’s “currently no support” for “Bring-your-own-key (BYOK) or bring-your-own-endpoint for additional rate limits,” so your Claude plan can’t help here. Our plan guide compares Pro and Ultra.
GitHub Copilot Pro+
Best for: people who already work in Copilot and want Opus 5.5 there.
GitHub’s model table marks Claude Opus 5.5 “Not included” on Copilot Pro and included on Pro+ and Max, in Copilot Chat, Copilot CLI, the cloud agent and the GitHub Copilot app. Every request spends AI credits at the model’s token price, “where 1 AI credit = $0.01 USD”: for Opus 5.5, $4 per million input tokens, $0.20 for cached input, $5 for cache writes and $20 per million output.
Price: Pro+ is $39 a month, and its plans page counts $70 of AI credits in it: $39 of base credits and a $31 flex allotment. Max is $100 a month, with $200.
Watch for: the flex part isn’t fixed: “Flex allotments may change over time.” Past the allowance, “additional usage is billed in GitHub AI Credits at the per-token rates,” the same list prices as Anthropic’s API. Our plan guide compares Copilot’s plans.
Kiro Pro
Best for: trying Opus 5.5 free for the rest of a month, then a fixed monthly bill.
Kiro’s models page offers Claude Opus 5.5 on Pro and every paid plan, priced in credits against its Auto mode: “a task that costs 10 credits on Auto would cost 20 credits on Opus 5.5.” Pro is $20 a month for 1,000 credits, add-on credits are $0.04 each, and the first time you upgrade, the rest of that month is free, with 500 trial credits.
Price: $20 a month; Pro+ is $40 for 2,000 credits.
Watch for: a credit isn’t a fixed amount of work: “More complex prompts, such as executing a spec task, typically cost more than 1 credit.” And the plan is for Kiro’s own apps and editors that support ACP, not other tools: “Use through third-party automation harnesses (such as OpenClaw) that route requests outside of Kiro’s native interfaces is not permitted.”
Cursor Premium
Best for: people who already work in Cursor.
Cursor is now “sold as part of The X Subscription, one subscription for Grok, Cursor, and X,” and its included usage covers its own models. “Third-party models like Claude, GPT, and Gemini aren’t part of the included pool. On Premium and above, they’re available in Cursor at the model’s API list price from the first request, paid from your prepaid credit,” say its pricing docs. Opus 5.5 is $4 per million input tokens and $20 per million output there.
Price: Premium is $30 a month, plus every Opus 5.5 token at list price.
Watch for: per token, Opus 5.5 costs what it does on the API, on top of the plan. Cursor’s pricing page says “Third Party model usage in Cursor is not included in the monthly prices shown and must be purchased separately.”
Enjoy
Best for: making a Claude plan go further before you pay for credits or Max.
Enjoy is a desktop app for Mac, Windows and Linux that runs Claude Code on your own Claude plan, beside Codex, Grok Build, Antigravity, Muse Code and local models through OpenCode. It doesn’t sell Opus or make it cheaper per token. What it does for the bill:
- Waits instead of paying. When Claude Code hits a limit, Enjoy waits until the limit resets, then continues where it left off, so work you leave running doesn’t need usage credits.
- Uses your other accounts. If you keep a second Claude account, Enjoy switches to it when one runs out. Having more than one is allowed; see our help page.
- Hands work to another lab. Switch to Codex on a ChatGPT plan, or another agent, in the same conversation, for the work that doesn’t need Opus.
- Says when a smaller plan would do. On Max, Enjoy tells you when four weeks of your Claude Code usage would have fit in a smaller Claude plan.
Price: Enjoy is one monthly subscription on top of your AI plans, and the first seven days are free, with a card to start. It includes your phone, any browser and one computer; each other computer costs extra each month. See pricing.
Watch for: it’s another bill, so it pays only if it saves you credits, an upgrade or time. In Enjoy, Antigravity runs only “the Gemini models Google offers your account,” so Opus 5.5 in Enjoy means Claude Code on Anthropic’s plan. Enjoy isn’t open source, and a waiting conversation resumes only once your computer is awake with Enjoy open.
When to stay on your current plan
- On Claude Pro, and rarely running out: stay. It’s the cheapest of Anthropic’s plans with Opus 5.5, and $200 a year if you pay up front.
- On Max or Ultra, and using most of it: stay. With the same value per dollar at every tier, moving down saves money only if you weren’t using the room, and buying the difference back in usage credits costs API rates.
- On Google AI Pro for Gemini: use Opus 5.5 in Antigravity before adding a Claude plan, until Google’s quota stops you most weeks.
- In Copilot or Cursor every day: if you reach for Opus only now and then, its list price there is simpler than a second subscription.
How we compared
We read Anthropic’s pricing page and API pricing docs, its help center articles on usage credits, monthly API credits and Claude Code on Pro, Max and Ultra, and Claude Code’s model and cost docs; Google’s Antigravity models and plans pages and its Google AI plans; GitHub’s Copilot plans, supported models and model prices; Kiro’s pricing and models pages; Cursor’s pricing page and model prices; Artificial Analysis’s comparison; and the public part of SemiAnalysis’s October 5 article, all on November 18, 2026. We quote each company’s own words for its limits, and we don’t estimate what a company hasn’t published, such as how much a plan’s limit holds. Prices change often, so we recheck this guide against its sources at least once a month, and the date at the foot of the page says when we last did.
15.Questions.
Is there a free way to use Claude Opus 5.5?
- Not on any free plan. Claude Free has no Opus or Claude Code, Antigravity’s free tier never had Opus 5.5 and loses Opus 4.6 on November 18, 2026, and Copilot Free can’t choose its model. The closest is Kiro’s trial. The first time you upgrade to Pro, “the rest of that month is free,” with 500 trial credits, and it needs a card.
What’s the cheapest subscription with Claude Opus 5.5?
- Claude Pro, at $20 a month or $200 a year, which runs Opus 5.5 by default in Claude Code. Google AI Pro, at $19.99 a month or $199.99 a year, has it in Antigravity on paid, non-trial plans, and Kiro Pro, at $20, has it at twice the credits of Kiro’s Auto mode.
Is the Claude API cheaper than Claude Pro for Opus?
- Only for light use. Opus 5.5 costs $4 per million input tokens and $20 per million output, and Artificial Analysis puts one of its benchmark tasks at $12.61 on Claude Code with Opus 5.5 at max effort, so two such tasks cost more than a month of Pro.
Is Claude Max cheaper than buying usage credits?
- Once you’d spend more than $80 a month on credits, yes. Usage credits “are billed at standard API rates,” Max is $100 for five times Pro’s usage, and SemiAnalysis measured a $200 Claude plan still at half its limit after $2,485 of Fable 5.1 at API prices.
Can I use Claude Opus 5.5 in GitHub Copilot or Cursor?
- Yes, on paid plans, at list price per token. Copilot marks Opus 5.5 “Not included” on Pro and includes it on Pro+, at $39 a month, from its AI credits. Cursor has it on Premium, at $30, and up, “at the model’s API list price from the first request, paid from your prepaid credit.”
Do Claude Max’s API credits pay for Claude Code?
- No. Anthropic says the monthly API credits on Max and Ultra “don’t cover interactive Claude Code sessions or extra usage after you hit your plan’s limits.” They do cover claude -p and the Agent SDK run with an API key from the Claude Console organization they’re linked to.