Quanta API Relay: Gemini 2.5 Flash Pricing, No Total Cap

01 · How does Quanta handle a long streaming answer without cutting it off?

A long answer is not the same thing as a stuck request. Quanta separates the two. Streams on Quanta have no total-time cap, so a response can keep running for as long as the model keeps producing tokens. The stream only ends after 300 seconds of upstream silence, meaning no data has arrived from the upstream provider for that period.

That distinction matters for teams running sustained workloads: report generation, large document review, code analysis and other jobs where the useful output arrives progressively. Instead of watching a wall-clock limit and wondering whether the request will be terminated, teams set a practical timeout and let the work finish.

Quanta publishes the surrounding limits as well, so engineers can plan capacity rather than discover constraints at the worst moment. Each key is limited to 120 requests per minute, each account can have 8 requests in flight, each request body can be 25 MB, and each account can hold 20 live keys. These are enforced limits, not suggestions, and Quanta states them plainly because a relay that hides its boundaries is hard to design around.

For teams, the result is a straightforward operating model: sustained streaming, visible concurrency, and a documented request ceiling. Quanta is built for heavy, sustained workloads, and the limits are the ones your team actually has to work with.

02 · What does Gemini 2.5 Flash cost on Quanta?

Quanta meters every call on real tokens, and input, output, cache writes and cache reads are each priced separately. The dashboard lists the tokens and the charge for every individual call, so a long stream can be reviewed line by line rather than inferred from a monthly total.

On Quanta, Gemini 2.5 Flash is priced per 1M tokens as follows:

• › Input: $0.2361
• › Output: $1.89
• › Cache write: $0.2361
• › Cache read: $0.0236

These are approximate USD figures, and charges are made in CNY.

As an example, at a 3:1 input to output ratio with no cache hits, Gemini 2.5 Flash on Quanta costs approximately $6.49 for 10 million tokens a month, or approximately $64.93 for 100 million tokens a month. These are worked examples, not quotes for your own workload: your real figure depends on your token mix, your cache behaviour and how much output your team requests.

Quanta is pay-as-you-go only. There is no monthly fee, no minimum spend and no subscription tier to justify. You fund a prepaid balance and draw it down as calls are metered. Top-up options are $1.39, $6.94, $13.89, $27.78, $69.44 and $138.89, so a team can start small and scale its balance as usage becomes predictable. New accounts also receive $0.50 in trial credit on sign-up, which is enough to confirm that the integration behaves as expected before a larger commitment.

03 · How do you set up Gemini 2.5 Flash through Quanta?

Gemini 2.5 Flash is used on Quanta over the native protocol, through Gemini CLI. It is not an OpenAI-compatible model, and Quanta does not describe it as one. If your team standardises on Gemini CLI, the path is direct: point the client at Quanta's API base address, authenticate with your Quanta key, and select the Gemini 2.5 Flash model.

Quanta serves OpenAI-compatible (/v1), Anthropic native and Gemini native protocols from one key and one balance. That does not mean every model works in every client. Gemini models use the native protocol only, through Gemini CLI. Claude models work only from the official Claude Code and Claude Desktop clients, with Claude Code 2.1.285 or newer. Codex, Cursor, Cherry Studio, Cline, OpenCode, OpenAI SDKs and cURL pair only with GPT, DeepSeek and the other non-Claude, non-Gemini models available on Quanta.

That clarity is deliberate. A team running several model families across several tools can still consolidate its API spend, but it will not spend an afternoon debugging a client and protocol combination that was never supported. Quanta publishes the pairing rules so integration work can be scheduled like any other engineering task.

04 · What does a team gain from one key and one balance?

Consolidation only helps if the accounting stays legible. Quanta charges each call on real tokens, separates input, output, cache writes and cache reads, and shows the tokens and charge for every call in the dashboard. For a team moving between GPT, DeepSeek, Gemini and Claude workloads, that means one balance to fund and one place to read usage.

There is no monthly fee and no minimum spend, so a quieter team is not paying for capacity it does not use. Unused balance can be refunded within 15 days of the top-up, which reduces the risk of funding a large balance too early.

Quanta also offers referral rewards: 19% of a referred account's top-ups, and 5% one level further. Rewards can be moved into the balance or withdrawn from $27.78.

For company billing, 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. Quanta does not run a support inbox; the dashboard and those pages are the route to the team. For teams that need cost visibility more than a flat fee, this is a practical middle ground.

05 · What happens to the prompts and answers sent through Quanta?

Data handling is stated directly. 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 useful when streaming long outputs. A response that runs for many minutes is not quietly retained as a copy of the answer. The input remains available for the stated operational reasons for a defined window, while the generated output is not kept.

Quanta also publishes per-model pricing for the rest of its catalogue, including GPT-6.1 Sol at $0.2083 input and $1.01 output per 1M tokens, and Claude Sonnet 5.5 at $1.25 input and $6.28 output per 1M tokens, with cache write and cache read prices listed separately for each. As an example, at a 3:1 input to output ratio with no cache hits, GPT-6.1 Sol on Quanta runs approximately $4.10 for 10 million tokens a month and approximately $40.97 for 100 million tokens a month. Again, these are worked examples for planning, not a quote of your own bill.

06 · Why start a Quanta account for long-running streams?

Because the operational questions are already answered. Streams have no total-time cap and end only after 300 seconds of upstream silence. Limits are published: 120 requests per minute per key, 8 requests in flight per account, 25 MB per request body, 20 live keys per account. Billing is metered on real tokens with a visible per-call breakdown, funded from a prepaid balance with no monthly fee and no minimum spend.

For Gemini 2.5 Flash specifically, Quanta keeps the setup honest: native protocol, Gemini CLI, one key and one balance alongside the rest of your supported models. For teams that want an API relay they can plan capacity and cost around, that is the point of Quanta.