generated: '2026-08-11'
method: searched
source: https://weaveapi.dev/pricing/
machine_readable_source: https://api.weaveapi.dev/api/pricing
docs: https://weaveapi.dev/docs/

model: prepaid-credits
currency: USD
plan_count: 2

summary: >-
  WeaveAPI sells one product — a prepaid credit balance spent across every supported model route — and
  publishes no per-seat or per-month subscription tier. The public pricing page describes two paths:
  self-serve "Pay as you go" and a "Account review" path for larger accounts that is a contact-support
  motion, not a priced tier. Neither path publishes a dollar rate on the public site; the docs, the
  pricing page and the models page all defer to the console Model Marketplace as the source of truth.
  That marketplace, however, has an UNAUTHENTICATED public projection at /api/pricing, which returns
  the full callable model list with its price ratios — captured verbatim in weaveapi-model-pricing.json.

plans:
  - name: Pay as you go
    id: pay-as-you-go
    price: null
    price_note: >-
      No fixed price. Prepaid credits are purchased in the console and drawn down per request; the
      rate depends on the model route selected. No published minimum top-up on the public site.
    currency: USD
    billing_period: prepaid
    self_serve: true
    highlighted: true
    included:
      - One shared account balance across all supported model routes
      - API keys and usage logs in console
      - Model rates shown in the live Model Marketplace
      - No separate provider dashboard required for WeaveAPI usage
      - New accounts include a $1 signup testing credit
    limits:
      rate_limits: Standard account limits (no numbers published)
      api_access: OpenAI-compatible /v1
      api_keys: Console managed
      usage_logs: Available in console
      support: Standard support handoff
    best_for: Testing, prototypes, apps

  - name: Account review
    id: account-review
    price: null
    price_note: Contact support. Not a priced tier — a review motion for larger usage.
    currency: USD
    billing_period: prepaid
    self_serve: false
    contact: support@weaveapi.dev
    included:
      - Everything in pay as you go
      - Account-limit review
      - Support handoff for operational questions
      - Usage planning discussion
      - Manual review for larger accounts
    limits:
      rate_limits: Account review required
      model_pricing: Marketplace plus account review
    best_for: Larger production workloads

signup_credit:
  amount: 1
  currency: USD
  note: New accounts include a $1 signup testing credit.

# Probed verbatim from the unauthenticated gateway config at /api/status
billing_configuration:
  method: probed
  source: https://api.weaveapi.dev/api/status
  quota_display_type: USD
  quota_per_unit: 500000
  display_in_currency: true
  price: 7.3
  stripe_unit_price: 8
  usd_exchange_rate: 7.3
  note: >-
    These are the gateway's own top-up/exchange settings as served publicly. `quota_per_unit` (500000)
    is the internal quota units per one display unit; `price` / `usd_exchange_rate` (7.3) and
    `stripe_unit_price` (8) are the top-up conversion settings. Payment is taken via Stripe.
  payment_processor: Stripe

# Price ratios, not dollar amounts — recorded exactly as the provider serves them.
pricing_model:
  scheme: ratio
  note: >-
    WeaveAPI runs the open-source New API gateway (confirmed by the x-new-api-version and
    x-oneapi-request-id response headers), which prices per model with a `model_ratio` multiplier for
    input tokens, a `completion_ratio` multiplier applied on top for output tokens, and a `cache_ratio`
    for cached input. NO DOLLAR FIGURE PER MODEL IS PUBLISHED ANYWHERE PUBLIC, and this artifact does
    not compute one — converting a ratio to a rate requires the gateway's base-rate constant, which
    WeaveAPI does not publish. The ratios below are recorded verbatim so the relative cost of routes is
    at least comparable.
  quota_type_0_means: token-metered (as opposed to per-call fixed price)
  groups:
    default: Default group
    vip: VIP group
  group_ratio:
    default: 1
    vip: 1
  group_note: >-
    A VIP group exists and is selectable, but its multiplier is 1 — identical to default. No model in
    the public catalog is restricted to it; every route lists enable_groups [default].

model_catalog:
  method: probed
  source: https://api.weaveapi.dev/api/pricing
  file: weaveapi-model-pricing.json
  fetched: '2026-08-11'
  model_count: 19
  pricing_version: a42d372ccf0b5dd13ecf71203521f9d2
  supported_endpoint:
    openai:
      path: /v1/chat/completions
      method: POST
  by_vendor:
    Alibaba Cloud (Qwen): 7
    OpenAI: 4
    Zhipu AI (GLM): 4
    Moonshot (Kimi): 3
    MiniMax: 1
    Anthropic: 0
    DeepSeek: 0
  coverage_gap:
    finding: >-
      WeaveAPI markets support for "OpenAI, Anthropic, DeepSeek, Qwen, Kimi, GLM and MiniMax" on its
      homepage, docs, models page and pricing page, and its own gateway registers Anthropic (vendor 9)
      and DeepSeek (vendor 2) in the vendor table — but the public callable catalog contains ZERO
      Anthropic routes and ZERO DeepSeek routes. Five of the seven advertised families are callable;
      two are not, at least not to an unauthenticated reader.
    caveat: >-
      This is the UNAUTHENTICATED projection of the catalog. WeaveAPI's docs state that the console and
      the authenticated /v1/models response are the source of truth and that access is account-specific,
      so an authenticated account may see routes this probe cannot. Recorded as an observed gap between
      the public marketing claim and the public catalog, not as a claim about what any account can call.
    also: >-
      All 19 routes declare supported_endpoint_types ["openai"] only. None declares "anthropic", even
      though the gateway exposes a live Anthropic-compatible POST /v1/messages route (401 unauthenticated)
      and the docs publish a Claude Code setup guide against it.
  models:
    - model: gpt-5.6-luna
      vendor: OpenAI
      model_ratio: 0.01
      completion_ratio: 6
      cache_ratio: 0.1
      groups: [default]
      endpoints: [openai]
    - model: gpt-5.6-terra
      vendor: OpenAI
      model_ratio: 0.1
      completion_ratio: 6
      cache_ratio: 0.1
      groups: [default]
      endpoints: [openai]
    - model: gpt-5.6-sol
      vendor: OpenAI
      model_ratio: 0.25
      completion_ratio: 6
      cache_ratio: 0.1
      groups: [default]
      endpoints: [openai]
    - model: gpt-5.5
      vendor: OpenAI
      model_ratio: 0.25
      completion_ratio: 6
      cache_ratio: 0.1
      groups: [default]
      endpoints: [openai]
    - model: qwen3.6-flash
      vendor: Alibaba Cloud
      model_ratio: 0.09
      completion_ratio: 6.6
      cache_ratio: 0.1
      groups: [default]
      endpoints: [openai]
    - model: qwen3-coder-next
      vendor: Alibaba Cloud
      model_ratio: 0.135
      completion_ratio: 4.777777777778
      cache_ratio: 0.1
      groups: [default]
      endpoints: [openai]
    - model: qwen3.5-plus
      vendor: Alibaba Cloud
      model_ratio: 0.18
      completion_ratio: 5.777777777778
      cache_ratio: 0.1
      groups: [default]
      endpoints: [openai]
    - model: qwen3.6-plus
      vendor: Alibaba Cloud
      model_ratio: 0.18
      completion_ratio: 6.7
      cache_ratio: 0.1
      groups: [default]
      endpoints: [openai]
    - model: qwen3.7-plus
      vendor: Alibaba Cloud
      model_ratio: 0.18
      completion_ratio: 3.888888888889
      cache_ratio: 0.1
      groups: [default]
      endpoints: [openai]
    - model: qwen3-coder-plus
      vendor: Alibaba Cloud
      model_ratio: 0.35
      completion_ratio: 4
      cache_ratio: 0.1
      groups: [default]
      endpoints: [openai]
    - model: qwen3-max-2026-01-23
      vendor: Alibaba Cloud
      model_ratio: 0.5
      completion_ratio: 5.2
      cache_ratio: 0.1
      groups: [default]
      endpoints: [openai]
    - model: glm-4.7
      vendor: Zhipu AI
      model_ratio: 0.27
      completion_ratio: 3.666666666667
      cache_ratio: 0.1
      groups: [default]
      endpoints: [openai]
    - model: glm-5.2
      vendor: Zhipu AI
      model_ratio: 0.3
      completion_ratio: 3
      cache_ratio: 0.166666666667
      groups: [default]
      endpoints: [openai]
    - model: glm-5
      vendor: Zhipu AI
      model_ratio: 0.45
      completion_ratio: 3.2
      cache_ratio: 0.1
      groups: [default]
      endpoints: [openai]
    - model: glm-5.1
      vendor: Zhipu AI
      model_ratio: 0.6
      completion_ratio: 3.333333333333
      cache_ratio: 0.1
      groups: [default]
      endpoints: [openai]
    - model: kimi-k2.5
      vendor: Moonshot
      model_ratio: 0.15
      completion_ratio: 3
      cache_ratio: 0.1
      groups: [default]
      endpoints: [openai]
    - model: kimi-k2.6
      vendor: Moonshot
      model_ratio: 0.2
      completion_ratio: 3
      cache_ratio: 0.1
      groups: [default]
      endpoints: [openai]
    - model: kimi-k3
      vendor: Moonshot
      model_ratio: 0.25
      completion_ratio: 3
      cache_ratio: 0.1
      groups: [default]
      endpoints: [openai]
    - model: MiniMax-M2.5
      vendor: MiniMax
      model_ratio: 0.135
      completion_ratio: 4.074074074074
      cache_ratio: 0.1
      groups: [default]
      endpoints: [openai]

refunds:
  policy_url: https://weaveapi.dev/refund/
  summary: >-
    Prepaid credits. Duplicate or incorrect charges are investigated; missing credits from a completed
    recharge are reconciled on request; credits already consumed for delivered usage are generally not
    refundable unless required by law or a confirmed billing error.

x-evidence:
  - url: https://weaveapi.dev/pricing/
    http_status: 200
  - url: https://api.weaveapi.dev/api/pricing
    http_status: 200
    content_type: application/json
  - url: https://api.weaveapi.dev/api/status
    http_status: 200
    content_type: application/json
  - url: https://weaveapi.dev/refund/
    http_status: 200