Cheap Kimi API: K3.1 Pro, K3 and K2.8 Preview Prices Compared

Kimi K3.1 Pro pricing: official API vs OpenRouter vs APIMaster

Compare the same model and billing mode. Kimi membership and Kimi Code subscriptions are separate from API credits.

The OpenRouter catalog was checked on the snapshot date; provider terms and account fees may differ. Official Batch is priced at 60% of standard rates, meaning 40% off. Caching needs a separate comparison: official Pro cached input is $0.20, with cache writes of $2 for a 5-minute TTL or $4 for a 1-hour TTL; output remains $10. The headline discount does not apply to every billing item.

Kimi API cost: how much can you save per task?

Calculate input_tokens × input_rate / 1,000,000 + billed_output_tokens × output_rate / 1,000,000. These are token-budget examples, not performance tests. Include billed reasoning output and every agent call.

At the second budget, 1,000 Pro tasks save $280 versus official standard rates. Excludes cache, tools, taxes and retries; this is not a bulk discount. Different models may consume different token counts. Compare cost per accepted task with the same acceptance tests.

Cheap Kimi K2.8 Preview or K3.1 Pro: which should you buy?

Choose Pro for the current Pro model at a discount. Its lowest rates are below K3, though existing applications may depend on K3's behavior. See the Pro capability guide and Pro versus Opus comparison.

Test K2.8 Preview for the smallest bill if it passes your extraction, formatting or coding tests. Use the exact ID kimi-k2.8-preview; the checked kimi-k2.8 and kimi-k2-8 IDs have no public routes. Low price does not establish Pro-equivalent quality.

Kimi's official platform closed K2-series access to new users on October 31, 2026 and plans retirement by the end of December 2026. APIMaster's listed Preview routes do not guarantee future lifespan. Keep the model configurable for migration.

How to get a discounted Kimi API key

Check the Pro card, register, fund your wallet and create an API key. Install the OpenAI Python SDK, then use your APIMaster key with https://apimaster.ai/v1 and kimi-k3.1-pro:

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["APIMASTER_API_KEY"],
    base_url="https://apimaster.ai/v1",
)
response = client.chat.completions.create(
    model="kimi-k3.1-pro",
    reasoning_effort="high",
    max_completion_tokens=4096,
    messages=[{"role": "user", "content": "Review this function for edge cases: def average(xs): return sum(xs) / len(xs)"}],
)
print(response.choices[0].message.content)
print(response.usage)

Start with a small request and inspect usage and billing. Pro needs room for reasoning tokens within its completion budget; check the finish reason for incomplete answers. Route-specific compatibility applies. This article checks prices and availability without a paid inference test.

How to choose a cheap Kimi route

Compare routes inside the same card and check the final quote. Some K3 routes exceed official standard rates. Fingerprints provide route-specific identity evidence; availability records describe test requests at a point in time. Neither guarantees every response or an uptime SLA; missing records do not count as passes. Highspeed has no public route at the snapshot.

For coding-client setup, follow the Kimi Code integration guide. An API discount does not include an official Kimi Code subscription.

FAQ

What is the cheapest Kimi API here?

Of these available models, Preview has the lowest listed input and output rates. Pro is cheaper than K3 at their checked minimum rates. This is not a claim of the internet's lowest price.

How do I get the Pro discount?

Open its model card, compare routes, register and create an API key. The maximum saving applies to the lowest checked route, not every route.

Is Pro cheaper than K3?

Yes at the checked lowest APIMaster rates and official standard rates. Completed-task cost still depends on token usage, retries and route choice.

Does a Kimi subscription include API credits?

Official API billing, Kimi membership and Kimi Code are separate. APIMaster usage is billed to APIMaster; it does not fund an official Kimi account.

Is K2.8 Preview still available?

It has public APIMaster routes at the snapshot. The official K2 retirement schedule is not a verified APIMaster shutdown date; keep a migration option.

Does the discount cover Highspeed and cache hits?

No. The headline compares ordinary uncached Pro input and output. Highspeed availability and cache billing require separate checks.

Get started: check Pro prices → register → create an API key → use https://apimaster.ai/v1 with kimi-k3.1-pro, up to 70% below official standard token rates.