Claude Sonnet 5 Pricing Guide for Quanta Teams: $1.25 Input

01 · What Claude Sonnet 5.5 costs on Quanta

Claude Sonnet 5.5 is a sensible starting point for teams that run heavy, continuous traffic through a coding assistant rather than occasional one-off prompts. On Quanta, every rate is published per 1M tokens and every category is metered separately, so you can see what you are paying for before the bill arrives.

• › Input: $1.25 per 1M tokens
• › Output: $6.28 per 1M tokens
• › Cache write: $1.57 per 1M tokens
• › Cache read: $0.0917 per 1M tokens

Those four numbers are the whole model. There is no bundled allowance, no seat charge, and no tier you have to guess about. Quanta is pay-as-you-go: the balance you hold is the balance you spend, and each call draws on real token counts.

For a straightforward illustration, consider an example using a 3:1 input-to-output ratio with no cache hits. At that mix, Claude Sonnet 5.5 works out to roughly $25.07 for 10 million tokens a month, or $250.69 for 100 million. These are examples, not quotes or commitments, but they give a planning baseline for a team that already knows roughly how many tokens its workflow consumes.

If your workload is lighter and you want the lower input rate, Claude Sonnet 5 is also on Quanta at $1.22 per 1M input tokens, $6.14 per 1M output tokens, $1.54 per 1M cache writes and $0.1069 per 1M cache reads. The two models sit close enough that the right choice usually comes down to how your team actually works, which is why both are worth comparing in the pricing view before you commit spend.

02 · Setting up Claude Code and Claude Desktop on Quanta

Claude models on Quanta are reached through the official Claude Code and Claude Desktop clients. That pairing is deliberate and narrow: if your team runs Claude Sonnet 5 or Claude Sonnet 5.5, these are the supported clients. Claude Code 2.1.285 or newer is required.

Claude models are not configured through an SDK, a pipeline, a command-line request, or a third-party editor, so there is no second integration path to maintain and no ambiguity about which tool a request came from. Teams that need a different editor can point it at other models listed on Quanta, but Claude traffic itself stays in the official clients.

The setup itself is short. Create an account on Quanta, add funds, generate a key, and enter your Quanta base address and key in the Claude client. Quanta speaks the Anthropic native protocol for Claude traffic, and the same key and the same balance also cover the OpenAI-compatible and Gemini native protocols used by other models on the platform. One credential, one balance, one dashboard.

Your first call does not need a large commitment. Sign-up includes $0.50 in trial credit, which is enough to confirm that the client connects, the model responds and the token meter is moving before you decide on a top-up amount.

03 · How Quanta handles sustained team traffic

Most relay friction appears when a team succeeds: more people, more parallel requests, more long-running streams. Quanta publishes its limits and enforces them, so capacity questions can be answered with numbers rather than guesswork.

• › 120 requests per minute per key
• › 8 requests in flight per account
• › 25 MB per request body
• › 20 live keys per account

The 8-request concurrency ceiling is the one to design around, because it applies to the account rather than to a single key. Twenty live keys give you room to separate workloads, environments or services, but they do not multiply the number of calls your account can have in flight at once.

Streaming behaves in a way that suits long sessions. Streams carry no total-time cap; they finish only after 300 seconds of upstream silence. In practice, a busy Claude Code session that is working through a large task is not cut off on an arbitrary timer. The limit that matters is silence from the upstream, not the length of your conversation.

Because the same key works across all three protocols, a team can route Claude Sonnet 5.5 through Claude Code and Claude Desktop while other services use the same credential. Quanta keeps the accounting in one place, which matters when the point of a relay is operational clarity rather than novelty.

04 · How pay-as-you-go billing works for Quanta teams

Quanta uses a prepaid balance with no monthly fee and no minimum spend. You fund the account, you use it, and you can stop. There are no passes or recurring commitments to unwind if a project slows down.

Top-ups are available at $1.39, $6.94, $13.89, $27.78, $69.44 and $138.89. The range covers a cautious first test and a sustained production load without forcing a large initial commitment. Unused balance can be refunded within 15 days of the top-up, which keeps the decision to fund low-risk for a new team.

Every call is metered on real tokens, with input, output, cache writes and cache reads priced separately. The dashboard lists the tokens and the charge for each individual call, so a reviewer can trace a specific request rather than reconcile an aggregate. For a team running many concurrent sessions, that granularity is the difference between a bill you accept and a bill you question.

Company payment runs through a bank transfer from a company account. Invoices and advance quotes are arranged through the Contact page or the Messages page of the dashboard, so procurement can request the paperwork it needs without hunting for a separate channel.

Referral rewards are also straightforward: 19% of a referred account's top-ups, plus 5% one level further. Rewards can be moved into the balance or withdrawn from $27.78. Teams that bring colleagues onto Quanta can convert that activity into working credit rather than treating it as a side note.

05 · What Quanta retains about your prompts

Data handling should be stated plainly. The text users send is kept for 30 days for abuse handling and billing disputes, and then deleted. Model answers are not stored.

That distinction is worth reading twice. The prompt side exists for a defined operational purpose and a defined window; the response side is not retained. For teams reviewing long Claude Code sessions, knowing which side of a request is held, for how long, and for what reason is more useful than a vague assurance.

06 · Choosing between Claude Sonnet 5 and Claude Sonnet 5.5

Both models are available on Quanta, both run through Claude Code and Claude Desktop, and both are billed per token with no plan structure. Claude Sonnet 5.5 is priced at $1.25 per 1M input tokens and $6.28 per 1M output tokens. Claude Sonnet 5 is priced at $1.22 per 1M input tokens and $6.14 per 1M output tokens.

A sensible first step for a team is to connect Claude Sonnet 5.5 in the official clients, run a small portion of real work against the $0.50 trial credit, and then read the per-call token breakdown in the Quanta dashboard. From there, top up at a level that matches observed usage and compare the resulting figures with the planning examples above.

Quanta is built for exactly this kind of workload: sustained traffic, concurrency and rate limits stated up front, invoices handled through the dashboard, and a price you can calculate before you spend. Open an account on Quanta and put Claude Sonnet 5.5 to work on a real task, with the numbers visible from the first call.