AI model price data — free JSON API

API

Free JSON API

One GET endpoint returns the whole snapshot. No key, no sign-up; responses are cacheable (about one hour) and cross-origin requests are allowed.

curl -s https://yysi.com/api/model-prices

Prefer a spreadsheet? Download the dated CSV snapshot: model-prices-snapshot.csv (the same checked rows, one line per price, formula-injection-safe).

Response shape (excerpt):

{
  "meta": {
    "name": "YYSI AI price snapshot",
    "updated": "2026-11-09",
    "checked": { "subscriptions": "2026-11-08", "tokens": "2026-11-09" }
  },
  "subscription_plans": [
    { "product_en": "Le Chat", "plan_en": "Pro", "price": 14.99,
      "currency": "USD", "ladder": "main", "checked": "2026-11-08",
      "source_url": "https://mistral.ai/pricing" }
  ],
  "api_token_prices": [
    { "model": "gpt-6.1-sol", "input_per_1m": 2.0, "output_per_1m": 10.0,
      "currency": "USD", "source_level": "official", "checked": "2026-11-09" }
  ],
  "yearly_billing": [ ... ],
  "entries": [ ... ]
}

Fields:

EMBED

Embed the price table

Drop this iframe into any page to show a compact live table of the mainstream monthly plans and per-token API prices. Chinese version: append ?lang=zh. Please keep the attribution line — it is what links back to the source.

<iframe src="https://yysi.com/embed/prices/"
        width="100%" height="800" style="border:0;border-radius:8px"
        loading="lazy" title="AI price table by YYSI"></iframe>

Live preview:

SNAPSHOT

Current snapshot

The numbers as checked on the dates above. The full directory, yearly ladder and all heavy tiers stay on the price comparison hub.

Full price comparison hub →

METHODOLOGY

Methodology & updates

Every row carries its own checked date and a source level: rows tagged “official page” were read on the vendor pricing page; rows tagged “third-party / media” quote trackers or reports and may lag.

Prices stay in the vendor list currency — no cross-currency conversion — and this is a comparison, not a ranking; there are no affiliate links.

Updates: prices are re-checked during routine editorial runs. Each verified re-check appends one dated snapshot to the changelog below, so trend lines accumulate over time.

Per-row observations: a dated observation records only the rows actually re-read that day (value, conditions, source). Diffs are computed only between retained observations — dates without a retained snapshot cannot be diffed and are never inferred.

Corrections: spot an error? Email [email protected] with the vendor, product, field, official link and your check date — corrections append to this changelog.

Quoting: cite the URL with the snapshot date and checked dates. The dataset is provided as-is with attribution appreciated; no dataset-redistribution license is granted and completeness is not promised.

• 2026-11-09 — First public snapshot published: 24 subscription entries, monthly/yearly ladders and 16 API token price rows (subscription plans checked 2026-11-08, token prices 2026-11-09).
• 2026-11-11 — Re-check #2: 6 of 16 API token rows re-read on vendor pricing pages (DeepSeek x2, OpenAI x3, Anthropic Sonnet 5.5) — all values unchanged against the 2026-11-09 baseline. Recording started: dates before 2026-11-09 have no per-row snapshots, so diffs are only ever computed between retained observations. 6 API token rows (details in rows)

Vendor / product	Item	Value	Conditions	Source
DeepSeek / deepseek-flash	input, cache miss, peak	2.0 CNY per 1M tokens	peak hours; cache hit 0.04; off-peak half	api-docs.deepseek.com
DeepSeek / deepseek-flash	output, peak	8.0 CNY per 1M tokens	peak hours; off-peak half	api-docs.deepseek.com
DeepSeek / deepseek-v4-pro	input, cache miss, peak	9.0 CNY per 1M tokens	peak hours; matches USD $1.32 (cache miss) on the official page	api-docs.deepseek.com
DeepSeek / deepseek-v4-pro	output, peak	27.0 CNY per 1M tokens	peak hours; matches USD $3.96 on the official page	api-docs.deepseek.com
OpenAI / gpt-6-astra	input / output, short context	10.0 / 50.0 USD per 1M tokens	standard tier	developers.openai.com
OpenAI / gpt-6.1-sol	input / output, short context	2.0 / 10.0 USD per 1M tokens	cached input $0.10; Batch = 50%; long context doubles	developers.openai.com
OpenAI / gpt-6-luna	input / output	0.10 / 0.50 USD per 1M tokens	standard tier	developers.openai.com
Anthropic / Claude Sonnet 5.5	input / output	2.0 / 10.0 USD per 1M tokens	cache write $2.50 (5m TTL) / $4 (1h TTL); cache read $0.10; Batch $1/$5	platform.claude.com

Change result: none detected in re-checked rows

YYSI, “AI subscription & price comparison”, snapshot 2026-11-11, https://yysi.com/subscriptions/ (accessed YYYY-MM-DD).