Quanta
EN
中文EnglishEspañolहिन्दीPortuguêsFrançais
Log inSign up

Built for high-volume workloads

At scale, every token deserves an exact line item

Quanta puts Claude, GPT, Gemini and other flagship models behind one API balance and prices each call on the usage the upstream actually reports — made for power users and teams running sustained, heavy workloads.

Sign up Browse the model catalog →
Monthly spend as volume grows (illustrative)Spend per month (log scale)
$0.10$1$10$100$1k$10k1M10M100M1BClaude Opus 5.5 · 1M · $3.46Claude Opus 5.5 · 10M · $34.58Claude Opus 5.5 · 100M · $345.83Claude Opus 5.5 · 1B · $3,458.33Claude Sonnet 5.5 · 1M · $2.51Claude Sonnet 5.5 · 10M · $25.07Claude Sonnet 5.5 · 100M · $250.69Claude Sonnet 5.5 · 1B · $2,506.94GPT-6.1 Sol · 1M · $0.3056GPT-6.1 Sol · 10M · $3.06GPT-6.1 Sol · 100M · $30.56GPT-6.1 Sol · 1B · $305.56Tokens per month (log scale)
  • Claude Opus 5.5
  • Claude Sonnet 5.5
  • GPT-6.1 Sol

Readout at four volumes

Monthly volumeClaude Opus 5.5Claude Sonnet 5.5GPT-6.1 Sol
1M$3.46$2.51$0.3056
10M$34.58$25.07$3.06
100M$345.83$250.69$30.56
1B$3,458.33$2,506.94$305.56

Computed from today's rate card, assuming a 3:1 input-to-output token mix and no cache hits. An illustration only — not any real customer's usage or bill.

Four things large accounts ask about

Concurrency, rate limits, credit and invoices — written as the system actually behaves.

01

Concurrency ceiling

Each account runs at most 8 requests in flight; anything beyond that gets an immediate 429 instead of waiting in a queue. The ceiling is the same for every account.

02

Rate limits per key

Each key accepts up to 120 requests in any 60-second window, with request bodies up to 25 MB. An account can hold 20 live keys, so services and environments can each have their own.

03

Credit and balance

Calls draw on a prepaid balance at the actual tokens of each request. New accounts start with $0.50 of trial credit; when the balance runs out the API answers 402 and resumes the moment you top up.

04

Invoices and reconciliation

The dashboard itemises the tokens and charge of every call, ready for month-end reconciliation. Invoices for top-ups are available — contact us with your username.

The model catalogue, vendor by vendor

Prices in $ per million tokens, input then output; all four price columns are on the pricing page.

Anthropic7 models

  • Claude Fable 5claude-fable-5$4.35 / $21.74
  • Claude Fable 5.1claude-fable-5-1$4.38 / $21.88
  • Claude Haiku 4.5claude-haiku-4-5-20251001$0.4167 / $1.97
  • Claude Opus 5claude-opus-5$1.97 / $9.83
  • Claude Opus 5.5claude-opus-5-5$1.79 / $8.46
  • Claude Sonnet 5claude-sonnet-5$1.22 / $6.14
  • Claude Sonnet 5.5claude-sonnet-5-5$1.25 / $6.28

OpenAI3 models

  • GPT-5.6 Solgpt-5.6-sol$0.4306 / $2.42
  • GPT-6 Astragpt-6-astra$0.8333 / $4.19
  • GPT-6.1 Solgpt-6.1-sol$0.1528 / $0.7639

Google4 models

  • Gemini 2.5 Flashgemini-2.5-flash$0.2361 / $1.89
  • Gemini 3.1 Progemini-3.1-pro$1.47 / $8.56
  • Gemini 3.8 Flashgemini-3.8-flash$1.35 / $6.10
  • Gemini 4 Argongemini-4-argon$1.72 / $8.28

DeepSeek2 models

  • DeepSeek V4 Flashdeepseek-v4-flash$0.2917 / $1.18
  • DeepSeek V4 Prodeepseek-v4-pro$1.44 / $4.32

Moonshot Kimi series1 models

  • Kimi K3kimi-k3$1.31 / $6.54

Alibaba Qwen series2 models

  • Qwen3.7 Plusqwen3.7-plus$0.1667 / $0.4861
  • Qwen3.8 Maxqwen3.8-max$0.8611 / $2.58

Zhipu GLM series1 models

  • GLM-5.3glm-5.3$0.6528 / $1.97

xAI Grok series2 models

  • Grok 4.6grok-4.6$0.6667 / $2.01
  • Grok 4.7grok-4.7$0.625 / $1.88

Tencent Hunyuan series1 models

  • Hunyuan 4hy4-preview$0.4167 / $1.32

Open the full pricing table →

Paying as a company?

For a bank transfer from a company account, invoices, or a quote confirmed in advance, contact us and we will fit in with your purchasing process.

Contactor write to [email protected]