Authenticity checks (model consistency and dilution) are not live yet; once they are, this section will compare the relay with the official API.便宜、直接的 GPT 上游接入。在一个平台探索模型 API、查看清晰价格，为你的下一个想法连接所需的工具。

Authenticity

Model consistency

Output distribution for the same requests on this relay and on the official API · higher is better

Once authenticity checks are live, this shows the distribution distance from the official API, capability-question pass rates side by side, and usage count drift. Routes that fail are greyed out on every leaderboard, with the reason shown in red on hover.

ⓘ Method: see the blog post “Did the model really get dumber?” Read

Speed

Measured by time to first token (seconds) and output speed (tokens / second)

Time to first token

Seconds until the first token · lower is better

0 relays

ⓘ Output speed: not yet available from the live API; shown once connected

Price

Cost per coding session

GPT 5.6 Luna · 2M input and 40k output tokens (¥) · split by token type · lower is better

1 relays

OutputInput

ⓘ USD prices are converted at 1 USD = 6.7698 CNY

Prices: cache hit, input and output

¥ / 1M tokens

Cache hitInputOutput

Cache

Coding requests repeat much of their context, so the cache hit rate largely decides what you actually pay. Drag the hit rate and the costs and ranking below are reweighted with cache prices.

Cache discount

Not provided

Cache hit — · input ¥0.012 (per 1M tokens)

Measured hit rate

Pending

Measured by replaying multi-turn sessions: cached-read tokens in usage as a share of all input

Session cost at 80% hit rate

¥0.027

¥0.027 without cache, 0% saved

Cache-weighted session cost

GPT 5.6 Luna · of 2M input tokens, the hit share is billed at the cache price and the rest at the input price, plus 40k output (¥) · lower is better

Rank 1 / 1

Cache hit rate80%Coding agents typically see 70%–90%

Cached inputUncached inputOutput

ⓘ Routes without a cache price are billed at the input price; cache write fees are not included

Stability

Availability

Request success rate, last 24 hours · higher is better

0 relays

Status of each model, last 24 hours

One cell per hour · green ok, yellow delayed, red failed

Model	Group	Availability	Last 24 hours	First token	Samples
Claude Opus 5.5	claude-kiro-撕破底价	77.4%		19.67 s	31
Kimi K3	国产模型-全网最低	77.4%		3.02 s	31
GPT 6 Astra	cheap-gpt-特惠	—	None yet	—	—
GPT 6.1 Sol	cheap-gpt-特惠	—	None yet	—	—
GPT 5.6 Luna	gpt-跌破底价	—	None yet	—	—
Claude Sonnet 5	claude-kiro-95%缓存	—	None yet	—	—
GPT 5.6 Sol	cheap-gpt-稳定	—	None yet	—	—
Claude Fable 5	claude-kiro-95%缓存	—	None yet	—	—
Claude Opus 5	claude-kiro-95%缓存	—	None yet	—	—
GPT 5.6 Terra	cheap-gpt-稳定	—	None yet	—	—
GPT 5.5	cheap-gpt-稳定	—	None yet	—	—

Service details

Minimum top-up20 USD

Sign-up bonus无

Support response工作日 4 小时内

Payment methodsAlipay, WeChat Pay, Bank card, Other

China invoiceSupported

International invoiceSupported

RefundsConditional

Promotions活动领取优惠

Strengths

价格极低，主打源头中转：官方宣传 0.08 倍费率、1 倍汇率，定位“GPT 源头中转，不限并发”。

Limitations

长上下文场景的可靠性需要加强

FAQ

How much does GPT 5.6 Luna cost on CheapAI?

Input ¥0.012 and output ¥0.072 per 1M tokens; at 2M input and 40k output tokens, a coding session costs about ¥0.027, an API rate of 0.0089.

How does the cache hit rate affect price?

CheapAI has not published a cache price for this route, so cache hits are billed at the input price.Measured hit rates will replace this assumption once available.

Which models does CheapAI support?

Folkbench currently lists 11: GPT 6 Astra, Claude Opus 5.5, GPT 6.1 Sol, Claude Sonnet 5, GPT 5.6 Sol, Claude Fable 5, Claude Opus 5, GPT 5.6 Terra, Kimi K3, GPT 5.5, GPT 5.6 Luna.

What does the data security level mean?

Folkbench rates relays from 0 to 4 based on public policies and verifiable evidence; CheapAI is currently level 3. The check covers leaks in responses and error messages only, and does not prove data retention or training policies.

How often is this data updated?

Availability and time to first token use a 24-hour window, and this page syncs live data every 15 minutes.