Claude Opus 5.5 Complete Guide: Pricing, Caching and Calling It on QCode
Anthropic released Claude Opus 5.5 (claude-opus-5-5) on 2026-09-22: 1M context, 128K max output, $4 input / $20 output per million tokens, cache reads at $0.20. As of 2026-10-19 it is still Anthropic's newest Opus: the official model overview (2026-10-13 snapshot) positions it for long-running agentic coding and knowledge work, with a reliable knowledge cutoff of June 2026, medium as the default effort, and retirement not sooner than 2027-09-22.
Updated 2026-10-22
Claude, GPT and Chinese models on one key, billed per token; 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 key points
List price per million tokens
$4 in, $20 out per million tokens; converted from the official price list, 20% below Opus 5 ($5 / $25).
Context / max output
1M tokens of context, up to 128K output tokens per request; the official model overview lists a reliable knowledge cutoff of June 2026.
Cache reads per million
Cache reads cost $0.20, which the official price list marks at 0.05x base input — 60% below Opus 5, converted from the list.
Claude Code default
Claude Code made claude-opus-5-5 its default Opus in version 2.1.280; the official model overview lists medium as the default effort for Opus 5.5.
What Claude Opus 5.5 is
As of 2026-10-19, Claude Opus 5.5 is the first model in Anthropic's new Claude 5.5 family. The official model overview positions it for long-running agentic coding and knowledge work, and says that if you are unsure which model to use, start with Claude Opus 5.5 for most workloads, and use Claude Fable 5.1 for demanding reasoning and long-horizon agentic work. Against the previous Opus 5, input and output prices are 20% lower and cache reads 60% lower, converted from the official price list.
Why it is worth a look now
Opus 5.5 is the first model in Anthropic's new Claude 5.5 family, released 2026-09-22; the launch post named Sonnet 5.5 and Haiku 5.5 as the rest of the family, and both have since released: Sonnet 5.5 landed on 2026-09-28, Haiku 5.5 on 2026-10-07. Claude Code has run claude-opus-5-5 as its default Opus since version 2.1.280. As of 2026-10-19, the official model overview (2026-10-13 snapshot) still lists Opus 5.5 as the newest Opus.
Public timeline
Anthropic releases Claude Opus 5.5, the first model in the new Claude 5.5 family, at 1M context / 128K output and $4 input / $20 output.
Anthropic releases Claude Sonnet 5.5; the Opus 5.5 launch post had already named Sonnet 5.5 and Haiku 5.5 as the rest of the family; Haiku 5.5 released on 2026-10-07.
Check date: the official model overview (2026-10-13 snapshot) says to start with Opus 5.5 and to use Fable 5.1 for demanding reasoning and long-horizon agentic work; the official deprecation table lists claude-opus-5-5 as Active, with retirement not sooner than 2027-09-22; Claude Code has shipped claude-opus-5-5 as its default Opus since version 2.1.280, and QCode can call it as of 2026-10-19.
Confirmed vs unverified
Confirmed
Confirmed by official sources: Anthropic released Claude Opus 5.5 on 2026-09-22 under the model ID claude-opus-5-5. The official model overview (2026-10-13 snapshot) lists 1M context, 128K max output, a reliable knowledge cutoff of June 2026, medium as the default effort and retirement not sooner than 2027-09-22, and the official deprecation table shows the model as Active. On the official price list: $4 input / $20 output per million tokens, cache reads $0.20, cache writes $5 (5 minutes) / $8 (1 hour); Batch $2 / $10; Fast mode $8 / $40 (per the official 2026-10-15 release notes, Fast mode for Opus 5.5 is out of research preview; for Opus 5 and Opus 4.8 it still is). QCode can call claude-opus-5-5 as of 2026-10-19.
Unverified · Anthropic's numbers only
Unverified: Anthropic says Opus 5.5 performs at the level of Claude Fable 5.1 on most work, costs 40% less to run than Opus 5, and reaches up to 2.5x the speed in Fast mode. All three are the company's own claims, unconfirmed by any third party; the official model overview only says to start with Claude Opus 5.5 for most workloads and to use Fable 5.1 for demanding reasoning and long-horizon agentic work, and it does not rank the two. No rumors or speculation beyond what Anthropic itself claimed are included. The launch post carries benchmarks, which are vendor-reported figures.
Line by line against Opus 5
Opus 5.5 vs Opus 5 (list price)
$4 input, $20 output per million tokens (Opus 5 is $5 / $25); cache reads $0.20 (Opus 5 is $0.50). Converted from the official price list: 20% less on input and output, 60% less on cache reads.
Anthropic's own cost claim vs the price-list cut
Anthropic says running costs on typical workloads are 40% lower than Opus 5 (the company's own testing); the cut on list unit prices is 20% (converted from the official price list). Two different bases — keep them apart.
How to choose, where to start
Here is the start-here line from the official model overview: if you are unsure which model to use, start with Claude Opus 5.5 for most workloads, and use Claude Fable 5.1 for demanding reasoning and long-horizon agentic work. If you would rather pay a higher unit price for faster output, look at Fast mode, priced at $8 / $40 on Opus 5.5.
Using it on QCode
As of 2026-10-19, claude-opus-5-5 is callable on QCode — put that ID in the model parameter of your request. The same key also works in Claude Code and the OpenAI Codex CLI: both clients point at QCode's compatible API, and switching models changes only that parameter. Claude Code has had it as the default Opus since version 2.1.280. Live pricing is on the models page. Want to use the new model for a few days first? Not ready for a monthly plan? Pay-as-you-go Packs start at $4 for a one-time quota you use within 7 days. Top-ups start at $5. If you already have a Personal Plan, buy a booster pack when the plan's quota runs short; Pay-as-you-go Packs are for accounts without an active plan.
FAQ
What does Opus 5.5 cost?
$4 input and $20 output per million tokens; cache reads $0.20, cache writes $5 (5 minutes) / $8 (1 hour); the Batch tier is $2 input / $10 output; Fast mode is $8 / $40. All official list prices, checked against the official price list fetched on 2026-10-19.
Anthropic says 40%, the price list 20% — is it 40% or 20% cheaper than Opus 5?
Converted from the official price list: input and output sit 20% below Opus 5 ($5 to $4, $25 to $20), and cache reads 60% below it ($0.50 to $0.20); both rows come from the official price list fetched on 2026-10-19. Anthropic separately says running costs on typical workloads fall 40% — that is a vendor-claimed running cost, not the same thing as the list-price cut.
How large is the context window, and the maximum output?
A 1M-token context window, with up to 128K tokens of output in a single request (official model docs, released 2026-09-22); the official model overview also lists a reliable knowledge cutoff of June 2026 and medium as the default effort.
Does Anthropic recommend it as the starting point?
The official model overview says it outright: if you are unsure which model to use, start with Claude Opus 5.5 for most workloads, and use Claude Fable 5.1 ($10 / $50) for demanding reasoning and long-horizon agentic work.
Can I call it on QCode?
Yes. The model ID is claude-opus-5-5, callable on QCode as of 2026-10-19, and the same key covers Claude Code and the Codex CLI, where switching means changing the model parameter. Claude Code has made it the default Opus since version 2.1.280. QCode bills per token, with per-model rates on the models page.
What is Fast mode?
Fast mode is Anthropic's option for faster output at a higher price per token. On Opus 5.5 it is priced at $8 / $40; the official 2026-10-15 release notes say it is out of research preview (Opus 5 and Opus 4.8 still are). The official figure is up to 2.5x the speed — that is the vendor's own number, with no third-party check. Prices checked 2026-10-19.
Sources
Sources: Anthropic's official model overview (2026-10-13 snapshot: model ID, 1M context / 128K output, default effort, knowledge cutoff, retirement date, the start-here line), the official price list (the Opus 5.5 and Opus 5 rows, the 0.05x cache-read footnote, the Batch and Fast mode tables), the official deprecation table (claude-opus-5-5 Active, retirement not sooner than 2027-09-22), the Opus 5.5 launch post ($4 / $20, the 20% and 60% cuts, the claimed 40% saving, Fable 5.1 parity and the claimed up to 2.5x speed), the Sonnet 5.5 launch post (family positioning and the coming-weeks line) and the Claude Code changelog (default Opus from 2.1.280). Checked 2026-10-19.
Try Claude Opus 5.5 right now
One key across the Claude and Codex lineup. As of 2026-10-19, Claude Opus 5.5 is available on QCode with claude-opus-5-5 - create an account, and it works as soon as you activate a Personal Plan or a Pay-as-you-go Pack.
Related reading
Claude Fable 5.2 Pricing: Official API Rates and Full Comparison (2026-11)
Item-by-item official input, output, cache-write and cache-read prices for Fable 5.2, Fable 5.1, Opus 5.5, Sonnet 5.5 and Haiku 5.5; batch and fast mode prices; worked cost examples; how to use it on QCode.
Claude Sonnet 5 guide
The middle tier at $2 input / $10 output — the value pick for everyday coding.
Kimi K3.1 Pro vs Claude Opus 5.5 (2026-11)
Official comparison of Kimi K3.1 Pro and Claude Opus 5.5: release dates, context, official max output, thinking behavior, official prices, open weights and vendor-reported benchmarks, plus switching between them on one QCode key.
Crypto payment guide
Nine coins, 56 networks, zero platform fee for AI coding API credits.
QCode is not affiliated with Anthropic. The specs and prices on this page come from Anthropic's official model overview, model docs, price list, deprecation table and launch post; the launch post carries benchmarks, which are vendor-reported figures. Live pricing is on the models page, and model availability is shown on /models, checked 2026-10-19.