Kimi K2.7 Code API pricing
If you mean Kimi K2.7 Code through Moonshot’s API, its published rate card lists $0.19 per million cache-hit input tokens, $0.95 per million cache-miss input tokens, and $4.00 per million output tokens for the standard model. These are usage charges, not a monthly subscription. (Kimi pricing)
| Model and source | Cache-hit input, per 1M tokens | Cache-miss input, per 1M tokens | Output, per 1M tokens |
|---|---|---|---|
kimi-k2.7-code — Moonshot API rate card |
$0.19 | $0.95 | $4.00 |
kimi-k2.7-code-highspeed — reported by a secondary source; not confirmed here on Moonshot’s rate card |
$0.38 | $1.90 | $8.00 |
The HighSpeed prices are reported by eesel’s pricing article, which says they are twice the standard rates. Treat them as secondary-source figures until you confirm them on the applicable rate card. Kimi’s API is pay-as-you-go, rather than an API subscription plan. (Kimi API pricing documentation)
Why you may see different K2.7 prices elsewhere
Provider matters. OpenRouter’s provider listing reports $0.60/M input and $3/M output for CoreWeave, and $0.95/M input and $4/M output for Moonshot AI. The listing identifies these as provider-specific prices. Its captured price information does not break input into cache-hit and cache-miss categories, so don’t assume those figures are directly comparable to every category on Moonshot’s rate card. (OpenRouter provider pricing)
Estimate your API cost
Use the rate for each input category separately:
Estimated cost = (cache-hit input tokens ÷ 1,000,000 × cache-hit rate) + (cache-miss input tokens ÷ 1,000,000 × cache-miss rate) + (billable output tokens ÷ 1,000,000 × output rate).
For example, a hypothetical standard-model workload with 1 million cache-hit input tokens, 1 million cache-miss input tokens, and 1 million billable output tokens would cost $5.14 before any other applicable charges: $0.19 + $0.95 + $4.00. Using the reported HighSpeed rates, the same token mix would cost $10.28. These examples assume the listed rates apply and that the stated output volume is the total billable output—not just the visible answer.
Another hypothetical: 100,000 cache-miss input tokens plus 10,000 billable output tokens on the standard model comes to $0.135: $0.095 for input + $0.04 for output. This assumes no cache hits and excludes other applicable charges.
Budget for reasoning, too. Kimi’s API documentation says K2.7 Code requires thinking to be enabled. (Kimi K2.7 Code quickstart) A secondary source reports that reasoning tokens are billed as output tokens, but the official material cited here does not confirm that billing detail. Until you verify how the applicable API reports and bills reasoning tokens, visible answer length alone is not enough for a defensible output estimate. Use total billable output from your usage data when available; the examples above treat their stated output volumes as billable output.
API usage is separate from a Kimi Code subscription
A Kimi Code subscription and pay-as-you-go API usage are separate products and charges. Don’t add a subscription price to an API budget unless you separately want both. (Kimi API pricing documentation)
Moonshot’s API recharge documentation states that a $1 minimum recharge is required to start API use and that a $5 voucher is granted when cumulative recharge reaches $5. Vouchers do not count toward the cumulative recharge total. These are account-funding terms, not K2.7 per-token prices. (Kimi recharge and rate limits)
Production budgeting notes
- Choose the model ID and provider route first. Moonshot’s standard-model rates differ by token category, while OpenRouter’s listing gives provider-specific rates without the same cache breakdown. Check the route you will actually use.
- Don’t assume every input token is cached. If you don’t know your workload’s cache-hit share, estimate a conservative case using the cache-miss rate, then update it using measured usage.
- Treat HighSpeed pricing as reported, not confirmed here. The secondary-source table lists higher rates; check the applicable rate card before using those figures for a production budget.
- Check throughput limits as well as cost. Kimi’s API documentation defines concurrency, requests per minute (RPM), tokens per minute (TPM), and tokens per day (TPD). Check your account’s current limits before launch; the cited documentation does not establish tier values for your account. (Kimi rate-limit documentation)
- Use measured token counts for a forecast. Price tables alone don’t establish your workload’s cache-hit rate, reliability, response quality, or monthly spend.
Bottom line: Moonshot’s listed standard Kimi K2.7 Code rates are $0.19/M cache-hit input, $0.95/M cache-miss input, and $4/M output. OpenRouter separately lists provider-specific rates, including CoreWeave at $0.60/M input and $3/M output. Keep provider route, token category, any subscription, and account recharge distinct when estimating cost.
Sources
What is Kimi K2.7 Code? Moonshot AI's coding model, explained
OpenRouter APIs — MoonshotAI: Kimi K2.7 Code provider pricing
Kimi K2.7 Code — Kimi API Platform