Claude Haiku 4.5 Pricing on Quanta: $0.4167 per 1M Input Tokens

01 · What Claude Haiku 4.5 costs on Quanta

Teams choosing a fast model for sustained, high-volume work usually want two things at once: a predictable per-token price and a service that will not surprise them at the end of the month. Quanta is built for exactly that shape of workload. Claude Haiku 4.5 on Quanta is billed per token, with no monthly fee, no minimum spend, and no plan tiers to negotiate. The published rate for claude-haiku-4-5-20251001 is $0.4167 per 1M input tokens, $1.97 per 1M output tokens, $0.4861 per 1M cache write tokens, and $0.0333 per 1M cache read tokens.

Those four numbers are the whole pricing story, and they are listed rather than buried in a calculator. Input, output, cache writes and cache reads are metered separately on the real tokens your calls consume, and the Quanta dashboard lists the tokens and the charge for every individual call. Prices are quoted in $ as approximate USD figures, while charges are made in CNY, so your finance team can reconcile the two without guessing.

For a sense of scale, the worked examples published for Quanta show Claude Sonnet 5.5 at approximately $25.07 for 10 million tokens a month and $250.69 for 100 million tokens a month, and Claude Opus 5.5 at approximately $34.58 for 10 million tokens and $345.83 for 100 million tokens, using a 3:1 input-to-output mix with no cache hits. These are examples, not quotes, but they show how the meter behaves when a team runs a large, continuous job.

02 · Setting up Claude Code 2.1.285 or newer for Quanta

Claude Haiku 4.5 is available through the Anthropic native protocol on Quanta, and Claude models connect only from the official Claude Code and Claude Desktop clients. There is no supported path through Cursor, Cherry Studio, Cline, OpenCode, Codex, an SDK, or cURL for any Claude model, so a setup that relies on those tools will not work for this one.

The practical sequence is short. Install or update Claude Code to 2.1.285 or newer, create a key in the Quanta dashboard, and point the client at the Quanta API base address using the Anthropic native protocol. From there, select the model identifier claude-haiku-4-5-20251001 and run your first request. If you prefer a desktop workflow, Claude Desktop is the other supported client for Claude Haiku 4.5.

One key and one balance cover all three protocols Quanta supports: OpenAI-compatible at /v1, Anthropic native, and Gemini native. That matters for teams that mix models by task rather than by preference. A Claude model runs in Claude Code or Claude Desktop; Gemini models use the native protocol through Gemini CLI; and Codex, Cursor, Cherry Studio, Cline, OpenCode, OpenAI SDKs and cURL pair with GPT, DeepSeek and the other non-Claude, non-Gemini models on Quanta. The same account carries the same prepaid balance across all of it, so adding a model does not mean opening a second billing relationship.

03 · Rate limits and concurrency for sustained runs

Heavy workloads fail in predictable ways: too many requests arrive at once, a single key is stretched past what the upstream will accept, or a long-running stream is cut off without warning. Quanta publishes and enforces its limits instead of leaving teams to discover them under load.

• › 120 requests per minute per key for request throughput.
• › 8 requests in flight per account for concurrency.
• › 25 MB per request body for payload size.
• › 20 live keys per account for separating workloads, services or environments.
• › No total-time cap on streams, which end only after 300 seconds of upstream silence.

For a team running continuous jobs, the concurrency ceiling is the number that shapes architecture. Eight requests in flight per account is a published, enforced boundary, so a queueing layer or a worker pool sized to it behaves predictably. Twenty live keys per account gives you room to separate a high-volume batch job from interactive work without multiplying your billing relationships. And because streams carry no total-time cap, a long generation is not interrupted simply for running long; it ends when the upstream goes quiet for 300 seconds.

04 · Prepaid balance, top-ups and invoices for company accounts

Quanta is pay-as-you-go only. There is no monthly fee and no minimum spend, so a team can start small and scale on measured usage rather than a committed allowance. New accounts receive $0.50 in trial credit on sign-up, enough to confirm the connection and see a first metered call in the dashboard before committing anything further.

Top-ups are offered at $1.39, $6.94, $13.89, $27.78, $69.44 and $138.89. That range covers both a quick verification run and a funded production balance, and the dashboard shows what each call consumed. Unused balance can be refunded within 15 days of the top-up, which keeps a small trial from becoming a stranded commitment.

For companies, payment runs by 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 get the paperwork it needs without a separate support channel. Referrals also carry real weight: a referred account's top-ups earn 19% in rewards, and 5% one level further. Rewards can be withdrawn from $27.78 or moved straight into the balance to cover usage.

05 · Prompt retention and the record behind each charge

Teams handling sensitive text need to know what is kept and for how long. On Quanta, the text you send is retained for 30 days for abuse handling and billing disputes, and then deleted. Model answers are not stored. The dashboard lists the tokens and the charge for every call, which means a disputed invoice can be checked against the actual request rather than a summary.

That combination — visible metering, published limits, a refund window and a defined retention period — is what makes Quanta practical for sustained team workloads. Claude Haiku 4.5 arrives at a stated $0.4167 per 1M input tokens and $1.97 per 1M output tokens, runs through Claude Code or Claude Desktop on the Anthropic native protocol, and sits on an account where one key and one balance span the OpenAI-compatible, Anthropic native and Gemini native protocols. Open a Quanta account, add the trial credit to your first test, and read the first line of usage in the dashboard before your team commits a production workload.