DeepSeek V4 Pro API relay pricing: One Quanta Account, 3 Protocols

01 · One balance, three protocols, and one set of published limits

When a team runs several model families, the usual administrative pattern is a separate vendor account for each one: separate sign-ups, separate balances, separate dashboards, and separate places to look when a call fails or a charge needs explaining. Quanta is built for a different arrangement. One Quanta account carries one key and one balance across three protocols: OpenAI-compatible at /v1, Anthropic native, and Gemini native. The API base address for all of them is the same, and the dashboard lists the tokens and the charge for every call.

That matters when the workload is heavy and sustained rather than occasional. Quanta publishes the limits it enforces instead of leaving them to be discovered under pressure: 120 requests per minute per key, 8 requests in flight per account, 25 MB per request body, and 20 live keys per account. Streams have no total-time cap and end only after 300 seconds of upstream silence. Nothing about that picture requires a monthly commitment. Quanta is pay-as-you-go, with a prepaid balance, no monthly fee, and no minimum spend.

For a team that has outgrown a patchwork of accounts, the practical question is simple: can one relationship carry the whole workload, with the numbers visible? On Quanta, the answer is yes.

02 · What a single Quanta account changes for a busy team

Consolidation is not only a tidier login screen. It changes how a team measures and defends its costs. Because every call is metered on real tokens, and because input, output, cache writes, and cache reads are priced separately, you can see the shape of the bill rather than a single blended figure. The dashboard shows the tokens and charge of every call, so a line item can be traced back to the request that produced it.

Consider an example: a team routes routine classification through one family and long-form generation through another, while a third family handles code work. Under separate vendor accounts, each of those streams arrives with its own top-up rhythm and its own reporting delay. Under one Quanta account, the same streams draw on one prepaid balance, and the same dashboard records all of them.

There is also a single support path. There is no support email address to hunt for; the team is reached through the Contact page or the Messages page of the dashboard. Company payment is arranged by bank transfer from a company account, and invoices and advance quotes are arranged through those same two pages. For a team that needs procurement paperwork alongside API traffic, keeping the commercial conversation and the technical usage in one place removes a recurring task.

03 · DeepSeek V4 Pro and the other families on one key

DeepSeek V4 Pro is priced on Quanta at $1.44 per 1M input tokens and $4.32 per 1M output tokens, with $0.0486 per 1M cache read tokens and no cache write price listed. It sits alongside a wide set of families on the same account, including DeepSeek V4 Flash, GPT-5.6 Sol, GPT-6 Astra, GPT-6.1 Sol, Qwen3.7 Plus, Qwen3.8 Max, Kimi K3, Hunyuan 4, Grok 4.6, Grok 4.7, GLM-5.3, the Gemini line, and the Claude line.

Because one key and one balance span all three protocols, a team can switch families without opening a second account. The trade-off is deliberate: Quanta makes the routing rules explicit rather than pretending that every model speaks the same dialect. The price list is per model family, and the protocol rules are stated per family, so the decision about which model to use is separated from the decision about where the account lives.

One account also keeps the concurrency picture legible. With 120 requests per minute per key and 8 requests in flight per account, a team can size its client configuration against a single published envelope. Twenty live keys per account give room for separating workloads or services, and 25 MB per request body is a concrete boundary to design around.

04 · Setting up clients without protocol mismatches

Quanta supports three protocols, and each model family connects through the right one. Getting this right at setup avoids most of the friction that would otherwise make a consolidated account look awkward.

• › Claude models work only from the official Claude Code and Claude Desktop clients, using Claude Code 2.1.285 or newer. That pairing is exclusive, so Claude traffic belongs in those two clients.
• › Gemini models use the native protocol only, through Gemini CLI. Gemini is not treated as OpenAI-compatible here.
• › Codex, Cursor, Cherry Studio, Cline, OpenCode, OpenAI SDKs, and cURL pair only with GPT, DeepSeek, and the other non-Claude, non-Gemini models listed in the Quanta catalogue, including Qwen3.7 Plus, Kimi K3, Hunyuan 4, Grok 4.7, and GLM-5.3.

DeepSeek V4 Pro therefore fits comfortably into the OpenAI-compatible lane, using the same /v1 base address as the GPT and Qwen models. If your team already drives several tools through an OpenAI-style client, the routing decision becomes a matter of model choice rather than account administration.

On data handling, Quanta states the position plainly: the text users send is kept for 30 days for abuse handling and billing disputes, then deleted, while model answers are not stored. Teams with heavy, sustained workloads can set expectations with that rule instead of an ambiguous policy.

05 · What the work actually costs on Quanta

Quanta quotes prices per 1M tokens, and the four components are separated so a workload can be modelled before it is run. The following worked examples use a 3:1 input-to-output ratio with no cache hits, and they are examples rather than quotes for any specific account.

• › 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 figures show why model selection and account structure are separate levers. A team can put most of its volume on a lower-priced family and reserve a premium family for the requests that need it, all inside the same Quanta balance. Because cache reads are priced on every listed family, and cache writes are priced on the families that support them, a team that reuses context can watch that effect in the per-call token breakdown on the dashboard.

Getting started is deliberately small. Sign-up includes $0.50 in trial credit, so a team can confirm the protocol pairing and the dashboard view before committing meaningful spend. Top-ups are offered at $1.39, $6.94, $13.89, $27.78, $69.44, and $138.89, which lets a team match a top-up to a measured month rather than buying a fixed package in advance.

06 · Invoices, refunds, and a clean way to bring the team in

Quanta is pay-as-you-go only, and the commercial terms are written down. There is no monthly fee and no minimum spend. Unused balance can be refunded within 15 days of the top-up, which is a straightforward position for a team that wants to trial a new model family without stranding funds.

For company accounts, payment is made by bank transfer from a company account, and invoices and advance quotes are arranged through the Contact page or the Messages page of the dashboard. That is also where the team handles questions, so procurement and usage questions do not need separate channels.

There is a further way to bring colleagues and collaborators onto the platform. Referral rewards are 19% of a referred account's top-ups and 5% one level further. Rewards can be moved into the balance or withdrawn once they reach $27.78, which is a practical route for teams that want colleagues using the same Quanta setup while the cost stays transparent.

The case for one account over several is not that a single vendor must be used for every task. It is that a team with heavy, sustained workloads should not have to administer several accounts to get access to several model families. Quanta puts the key, the balance, the concurrency limits, the token-level charges, and the invoice conversation in one place, with the prices listed in full. If that is the arrangement your team has been assembling by hand, open a Quanta account, add $0.50 of trial credit, and route the first model family through it.