Update 2026-09-30: we corrected how Claude orders work. Pro, Max 5x and Max 20x are added to your own account from a claude.ai cookie export checked before you pay; a ready-made Max is hand-delivered after you verify a delivery email, and a ready-made Pro needs its Organization ID submitted after it arrives. Details under “How ordering works here” below.
Get Claude at ClaudeMax: Claude Pro ¥249 · Max 5x ¥999 · Max 20x ¥1,799 on your own account (from a claude.ai cookie export checked before you pay, usually 5–30 minutes) · Claude Pro ready-made ¥329. Pay in USDT (TRC20 or BEP20); prices are whatever the pricing page shows when you order.
Opus 5.5 and Sonnet 5.5 at a glance
| Item | Opus 5.5 | Sonnet 5.5 | For reference: Fable 5.1 |
|---|---|---|---|
| Released | 2026-09-22 | 2026-09-28 | 2026-09-01 |
| Positioning | Long-running agentic coding and knowledge work; the default choice for most tasks | A faster, cheaper companion to Opus 5.5: everyday tasks, bug fixes, documents and spreadsheets | The hardest reasoning and long-horizon agentic work |
| API price per 1M tokens (input / output) | $4 / $20 | $2 / $10 | $10 / $50 |
| Cache reads per 1M tokens | $0.20 | $0.10 | — |
| Context / max output | All three: 1M-token context, 128K max output, adaptive thinking, June 2026 knowledge cutoff | ||
| API default effort | medium | high | high |
| What Anthropic claims | Fable 5.1-level on most work; 40% cheaper than Opus 5 on typical workloads; 30%+ faster output | 30%+ faster than Sonnet 5; up to 30% less per task | — |
Anthropic says Haiku 5.5 will follow in the coming weeks.
What changed for subscribers
- Opus 5.5 is the default. Anthropic's Claude Code lead Cat Wu confirmed Opus 5.5 is the default model in the Claude app (including Cowork) and Claude Code for Pro, Max and Team, at medium effort. Pro users now start on Opus 5.5.
- Higher five-hour limits. Anthropic says it is increasing five-hour usage limits on Pro, Max, Team and seat-based Enterprise plans, without a figure; outlets reported 20%. The weekly limit is separate.
- A saved limit reset. Subscribers get one rate-limit reset they can keep and use whenever they choose — useful on a crunch day.
- The same allowance goes further. Opus 5.5 costs 40% less than Opus 5 on typical workloads and Sonnet 5.5 up to 30% less per task. Limits are consumed by usage, so cheaper models stretch the same allowance.
| Model | Free | Pro | Max 5x / 20x |
|---|---|---|---|
| Sonnet, Haiku | Yes | Yes | Yes |
| Opus (now Opus 5.5, the default) | No | Yes | Yes |
| Fable 5.1 | No | Usage credits only | Up to 50% of the weekly allowance |
| Claude Code | No | Yes, sharing the plan allowance | Yes, 5× / 20× Pro's allowance |
Opus 5.5, Sonnet 5.5 or Fable 5.1
Anthropic's model docs are direct: start with Opus 5.5 for most workloads, and use Fable 5.1 for the hardest reasoning and long-horizon agentic work, or when Opus 5.5 at higher effort falls short. By task:
| Task | Pick | Why |
|---|---|---|
| Everyday questions, email, documents, slides, spreadsheets | Sonnet 5.5 | Exactly what Anthropic positions it for; faster and lighter on limits |
| Well-scoped coding and bug fixes | Sonnet 5.5 | Same reason; cheaper when running several agents in parallel |
| Work that needs judgment; long agentic coding sessions | Opus 5.5 | Anthropic's default choice; Fable 5.1-level on most work |
| Problems Opus 5.5 can't crack at high effort | Fable 5.1 | Up to 50% of the weekly allowance on Max; usage credits on Pro |
Vendor benchmarks
These are the figures on Anthropic's launch pages (vendor-run, not independent), useful for seeing where each model is strong:
| Benchmark | Opus 5.5 | Sonnet 5.5 |
|---|---|---|
| Terminal-Bench 4.0 | 66.4% | 70.6% |
| FrontierCode v1.1 | 54.4% | 46.2% |
| CursorBench 4.0 | 57.8% | 55.5% |
| OSWorld 2.1 | 81.8% | 80.1% |
| GDPval-AA v2.1 | 1846 Elo | 1844 Elo |
| AutomationBench | 40.0% | — |
Sonnet 5.5 scores higher on Terminal-Bench, consistent with its positioning for well-scoped terminal and coding tasks; for long chains of judgment, Opus 5.5 remains the pick.
Using them in Claude Code
- Update Claude Code first:
claude update. Pro, Max and Team start on Opus 5.5. - Switch to Sonnet 5.5 to save allowance with
/model; Fable 5.1 is in the same menu (Max includes a Fable allowance). - Watch for
ANTHROPIC_API_KEY: if that environment variable is set, Claude Code bills the API instead of your subscription.
Pro or Max?
| How you use Claude | Pick | Our price |
|---|---|---|
| Writing, research, occasional Claude Code | Pro: Opus 5.5 is now the default and five-hour limits went up | ¥249 ≈ 37.11 USDT on your account; ready-made ¥329 |
| Claude Code daily, limits hit 3+ times a week | Max 5x: 5× Pro per session, includes Fable | ¥999 on your account; ready-made sold out for now |
| Agents running all day, heavy Fable 5.1 use | Max 20x: 20× Pro per session | ¥1,799 on your account; ready-made sold out for now |
An occasional overrun is what the saved reset and usage credits are for; upgrade when it happens every week. For the Max tiers, see Claude Max 5x or 20x.
How ordering works here
- On your own account: Claude Pro, Max 5x and Max 20x are added from a claude.ai cookie export (Cookie-Editor JSON) that you paste on the order page. It is checked before you pay and cleared after delivery, usually within 5–30 minutes, and the account stays yours. For Max the account must sign in normally and not be banned; a Free account that has never paid by card can't buy Max, and brand-new accounts wait 3–15 days before Max 20x. Expired or active Pro and Max accounts can renew or upgrade.
- Ready-made account: if you have no account or yours doesn't meet the top-up conditions. After paying you enter and verify a delivery email; a ready-made Max is then hand-delivered (the order page shows 30 minutes to about 4 hours), and a ready-made Pro arrives by email, after which you submit its Organization ID on the order page. Renewals cost the same as a top-up. Full guide: Ready-made Claude Pro and Max accounts. Ready-made Max 5x / Max 20x is sold out for now; for Max 5x / Max 20x, use the top-up on your own account (Max 5x / Max 20x) under the rules above.
- Pay in USDT (TRC20 or BEP20); contact the bottom-right live chat promptly if anything goes wrong.