Claude Code pricing for teams: one Quanta balance, 20 live keys

01 · One Quanta balance for a whole engineering team

Moving a team onto a new model provider usually means a stack of separate accounts, separate balances and separate invoices to reconcile at the end of the month. Quanta removes that fragmentation. Every developer in your team draws on one prepaid balance, and one key works across the OpenAI-compatible, Anthropic native and Gemini native protocols. There is no monthly fee and no minimum spend, so a small pilot and a large sustained workload are billed on the same terms: pay-as-you-go, metered on the tokens your calls actually use.

That model is a natural fit for engineering teams running Claude Code through long, continuous sessions. Quanta is built for teams and heavy, sustained workloads, and the account-level numbers are published rather than left to be discovered under pressure. You can see the limits before you route traffic, and you can see the charge for every individual call afterwards.

For a team that wants to start small, Quanta offers $0.50 in trial credit on sign-up. Top-ups are available at $1.39, $6.94, $13.89, $27.78, $69.44 and $138.89, so you can size the first deposit to the first week of usage instead of committing to a large balance. Unused balance can be refunded within 15 days of the top-up.

02 · Concurrency and rate limits you can plan around

The most common source of friction in a shared AI account is not the model catalogue but the throttling. Quanta publishes and enforces the following limits on every account:

• › 120 requests per minute per key
• › 8 requests in flight per account
• › 25 MB per request body
• › 20 live keys per account
• › Streams with no total-time cap, ending only after 300 seconds of upstream silence

Read those together and the shape of a team deployment becomes clear. Twenty live keys give you room to separate projects, services and individuals without creating a new account for each. The per-key rate limit means one busy service does not have to monopolise the account, while the in-flight cap of 8 requests applies at the account level, so it is worth watching that number as your parallel Claude Code sessions grow.

The stream behaviour matters for long-running work in particular. Because streams carry no total-time cap, a long agentic task is not cut off simply for taking a long time; a stream ends only when upstream has been silent for 300 seconds. For teams running sustained sessions across many engineers, Quanta removes the guesswork from capacity planning: the ceilings are stated, and they are the same ceilings you will see in the dashboard.

03 · What Claude Code usage costs on Quanta

Quanta prices per 1M tokens and separates input, output, cache writes and cache reads, so you can read the cost of a workload rather than a blended average. All listed prices are approximate USD figures, and charges are made in CNY. The dashboard lists the tokens and the charge for every call, which is what makes a shared balance auditable at the level of individual requests.

Here are worked examples, at a 3:1 input to output ratio with no cache hits:

• › Claude Opus 5.5: 10 million tokens a month ≈ $34.58; 100 million ≈ $345.83
• › Claude Sonnet 5.5: 10 million tokens a month ≈ $25.07; 100 million ≈ $250.69
• › GPT-6.1 Sol: 10 million tokens a month ≈ $4.10; 100 million ≈ $40.97
• › Gemini 2.5 Flash: 10 million tokens a month ≈ $6.49; 100 million ≈ $64.93

These are examples, not quotes for your team, but they show the range Quanta covers. A team running Claude Opus 5.5 through Claude Code for a large share of its coding work can budget against a per-million figure and then check the actual number afterwards. If part of the workload is lighter, Claude Haiku 4.5 is priced at $0.4167 per 1M input tokens and $1.97 per 1M output tokens, with cache writes at $0.4861 and cache reads at $0.0333.

Because the balance is prepaid and there is no minimum spend, the same account can carry a mixed workload: Claude Sonnet 5.5 for routine work, Claude Opus 5.5 for the harder problems, and a non-Claude model for tasks where a lighter price point makes more sense. One balance, one set of account limits, one invoice trail.

04 · Keys, company payment and invoices without the monthly fee

Getting a team live on Quanta takes a short list of administrative steps. Create the account, add credit, and issue up to 20 live keys. The API base address is a single endpoint for all three protocols, so the same key and the same balance serve Claude, GPT, DeepSeek, Gemini and the other supported models.

If your company pays by bank transfer, invoices and advance quotes are arranged through the Contact page or the Messages page of the dashboard. That is the route for procurement: you can request the paperwork your finance team expects while the engineers keep working from the same balance. There is no monthly subscription to explain, no seat count to reconcile and no minimum spend to justify.

For teams that grow by referral, Quanta pays 19% of a referred account's top-ups and 5% one level further. Rewards can be withdrawn from $27.78 or moved into the balance, which lets you keep them inside the account that is doing the work.

05 · Pairing Claude Code with the rest of your stack

Client pairing is the detail worth getting right before a rollout. Claude models work only from the official Claude Code and Claude Desktop clients, with Claude Code 2.1.285 or newer. Gemini models use the native protocol only, through Gemini CLI. Codex, Cursor, Cherry Studio, Cline, OpenCode, OpenAI SDKs and cURL pair only with GPT, DeepSeek and the other non-Claude, non-Gemini models on Quanta.

That means a mixed stack is straightforward rather than improvised. Your Claude Code engineers point at the Anthropic native protocol and use Claude models such as Claude Fable 5.1, Claude Opus 5.5 or Claude Sonnet 5.5. Gemini CLI connects to Gemini models such as Gemini 3.1 Pro or Gemini 2.5 Flash through the native protocol. Everything else in your tooling can use the OpenAI-compatible path with GPT, DeepSeek, Qwen, Kimi, Grok, GLM or Hunyuan models, all drawing on the same balance.

On data handling, the text users send is kept for 30 days for abuse handling and billing disputes, then deleted. Model answers are not stored.

06 · A rollout checklist for the first week

To move a whole engineering team across in a controlled way, start with a short sequence. Sign up and use the $0.50 trial credit to confirm connectivity from one machine. Deposit enough for the first week of real traffic rather than a large sum, and keep the refund window in mind: unused balance can be refunded within 15 days of the top-up.

Then issue keys by project rather than by person, keeping the 20 live key ceiling and the 8 in-flight request account limit in view. Configure Claude Code 2.1.285 or newer for Claude work, Gemini CLI for Gemini models, and the OpenAI-compatible path for the rest of your tooling. Watch the dashboard as you go: every call shows its tokens and its charge, so a surprising number can be traced to a specific request rather than a vague monthly total.

Finally, if the finance team needs documentation, arrange the bank transfer, invoice and advance quote through the Contact page or the Messages page of the dashboard. There is nothing else to set up, no plan to buy and no monthly commitment to unwind later.

Quanta gives a team one prepaid balance, published concurrency and rate limits, token-level metering and company-ready invoicing. For engineering groups standardising on Claude Code, that is a straightforward account to open and a straightforward set of numbers to plan against.