Moonshot AI
Operational· checked every few minutes

Kimi K3.1 Pro API

The Kimi K3.1 Pro API for coding agents, 3D work and financial research, 60% below Moonshot's list price

Price per 1M tokens60%off
Input
$0.80$2.00
Output
$4.00$10.00
Cached input
$0.08
Get your API keyTry it in the playground
Model IDkimi-k3.1-pro
chat.py
from openai import OpenAI
client = OpenAI(
    base_url="https://api.zurelay.com/v1",
    api_key="YOUR_ZURELAY_KEY",
)
stream = client.chat.completions.create(
    model="kimi-k3.1-pro",
    messages=[{"role": "user", "content": "Explain quicksort in two sentences."}],
    stream=True,
)
for chunk in stream:
    if chunk.choices:
        print(chunk.choices[0].delta.content or "", end="")

Base URL https://api.zurelay.com/v1. Works with any OpenAI SDK. Chat API docs

Context window
1M tokens
Max output
32K tokens
Reasoning
low · high · max
Input
Text, Image
Released
Oct 24, 2026
Status
Operational

Pricing

Kimi K3.1 Pro API pricing. A fraction of the list price.

Pay as you go from prepaid credit, with no subscription. Top up from $10. Requests that fail are never billed.

RateZurelayMoonshot AI list
Input
per 1M tokens
$0.80$2.00
Output
per 1M tokens
$4.00$10.00
Cached input
per 1M tokens
$0.08—

One rate at every prompt length, up to the full 1M context window. Streaming, tool calls and structured outputs cost nothing extra.

Every Kimi model and price
Savings calculator

Move the sliders to your monthly usage.

Input tokens per month50M
Output tokens per month10M
Zurelay
$80.00
Moonshot AI list price
$200.00
You save every month$120.00

$1,440 a year

Overview

What is Kimi K3.1 Pro?

Kimi K3.1 Pro is Moonshot AI's newest flagship and the first model of its K3.1 series: Kimi K3 trained much further with reinforcement learning, with native vision and a 1M-token context window. The Zurelay Kimi K3.1 Pro API serves the same model through one OpenAI-compatible endpoint at 60% below Moonshot's list price.

Moonshot AI released Kimi K3.1 Pro in October 2026 as the first model of its K3.1 series, together with Kimi K3.1 Pro Highspeed, the same model at two to three times the speed. K3.1 Pro is a post-trained version of Kimi K3, the 2.8-trillion-parameter open-weight model, with far more post-training compute spent on reinforcement learning across hundreds of domains, and Moonshot reports gains in every one of them. The full weights are due by November 4, 2026.

Moonshot reports 65.7% on Terminal-Bench 4.0, ahead of GPT-6 Astra at 58.2% and just behind Claude Opus 5.5 at 66.4%, and 61.2% on its own long-horizon Kimi Code Bench 3.0, up from K3's 55.95%. It scores 92.7% on BenchCAD with tools, rebuilding 3D objects as CAD code, and a state-of-the-art 64.4 on OfficeQA Pro v2, a test of analysis over real financial documents. On Moonshot's hallucination stress test, its claim-level error rate fell from 23.4% on K3 to 8.3%.

It covers coding from quick fixes to whole projects, builds games and web experiences in Unity, Unreal Engine, Godot and Three.js with more polished design, writes and runs Blender and CAD scripts to build parametric 3D models, and carries financial modeling and research work from start to finish. For the open-weight release Moonshot removed cyber-offense data from training, so its exploit-writing ability stays below what the same training would otherwise produce.

Thinking is always on. The API takes a reasoning_effort of low, high or max, with max as the default, and fixes temperature at 1.0 and top_p at 0.95. It reads text and images, calls tools, including tool_choice required, and holds 1,048,576 tokens of context. Moonshot lets one answer run to 131,072 tokens by default, but on Zurelay a kimi-k3.1-pro answer stops at 32,000 output tokens, reasoning included; Kimi K3.1 Pro Highspeed writes past that. On Zurelay you call it at https://api.zurelay.com/v1 with the model ID kimi-k3.1-pro, for $0.80 per 1M input tokens and $4.00 per 1M output tokens against Moonshot's list price of $2.00 and $10.00.

Strengths

Where Kimi K3.1 Pro shines. And what teams build with it.

01

Agentic coding

65.7% on Terminal-Bench 4.0 in Moonshot's tests, ahead of GPT-6 Astra and within a point of Claude Opus 5.5, and well ahead of K3 on long-horizon tasks.

02

3D and CAD

Turns a brief, a sketch or a reference image into a 3D model, and writes Blender and CAD scripts for parametric parts: 92.7% on BenchCAD with tools.

03

Financial work

State of the art on OfficeQA Pro v2, and ahead of K3 by 7.2 points on Moonshot's financial modeling tests and 11.6 on its financial research tests.

04

Fewer made-up facts

On Moonshot's hallucination stress test, the share of false claims fell from 23.4% on K3 to 8.3%.

Use cases

  • Coding agents

    Run long refactors, feature builds and test loops over a large repository in one 1M-token session, with the model driving terminal tools.

  • Games and interactive web

    Build Three.js scenes, landing pages and Unity or Godot prototypes, then send screenshots back so it can refine the result.

  • 3D modeling scripts

    Have it write and fix Blender or CAD scripts for mechanical parts, furniture, products with options, or architectural models.

  • Financial analysis

    Build linked spreadsheet models, check figures across filings and write research grounded in the documents you send.

Get started

Call Kimi K3.1 Pro in three steps.

No waitlist and no new SDK. If your code already talks to OpenAI, it already talks to Zurelay.

  1. 1

    Create a key

    Sign up, add credit and create an API key. Set a monthly budget or a rate limit per key if you like.

    Get your API key
  2. 2

    Point your SDK at Zurelay

    Change the base URL. Everything else in your code stays the same.

    https://api.zurelay.com/v1
  3. 3

    Use kimi-k3.1-pro

    Send chat completions as usual. Streaming, tool calls and usage reporting work as you expect.

    kimi-k3.1-pro

Works with the tools you already use

Compare

Similar models. All on the same key.

18%off

Kimi K3.1 Pro Highspeed

$3.30 / $16.50 per 1M tokens

Pick Kimi K3.1 Pro Highspeed for the same model at about twice the speed, for interactive apps or answers over 32,000 tokens, at a higher price.

Kimi K3.1 Pro Highspeed API
58%off

Kimi K3

$1.27 / $6.37 per 1M tokens

Pick Kimi K3 if your prompts are tuned on it. K3.1 Pro is its post-trained successor, with a lower list price.

Kimi K3 API
65%off

Claude Opus 5.5

$1.40 / $7.00 per 1M tokens

Pick Claude Opus 5.5 for the strongest agentic coding: Moonshot puts K3.1 Pro within a point of it on Terminal-Bench 4.0.

Claude Opus 5.5 API

FAQ

Kimi K3.1 Pro API questions.

Anything else? Ask our team and we’ll answer by email.

How much does the Kimi K3.1 Pro API cost on Zurelay?

Kimi K3.1 Pro costs $0.80 per 1M input tokens and $4.00 per 1M output tokens on Zurelay, with cached input at $0.08. Moonshot's list price is $2.00 input and $10.00 output, so you save 60%. One rate applies at every prompt length, up to the full 1M-token window.

Is there a cheaper Kimi K3.1 API than Moonshot's own?

Yes. Zurelay serves the same kimi-k3.1-pro model at 60% below Moonshot's list price. You top up prepaid credit and pay only for the tokens you use, with no subscription.

How do I call the Kimi K3.1 Pro API with the OpenAI SDK?

Install the official OpenAI SDK, set the base URL to https://api.zurelay.com/v1 and use your Zurelay API key. Then pass model "kimi-k3.1-pro" to chat.completions.create. Leave temperature and top_p unset, since Moonshot fixes them, and use reasoning_effort to pick low, high or max.

How is Kimi K3.1 Pro different from Kimi K3?

Kimi K3.1 Pro is Kimi K3 post-trained with much more reinforcement learning, so it has the same architecture, 1M-token context and vision. Moonshot reports clear gains in coding, 3D modeling and finance, and a much lower hallucination rate. Its list price is lower than K3's.

Should I use Kimi K3.1 Pro or Kimi K3.1 Pro Highspeed?

They are the same model, and Moonshot says they are equally capable. Highspeed writes two to three times faster: in our tests about 97 tokens a second against about 46 for K3.1 Pro. Pick Highspeed when someone is waiting on the answer, and K3.1 Pro for background jobs, where its lower price matters more than speed.

What is the Kimi K3.1 Pro context window?

Kimi K3.1 Pro holds 1,048,576 tokens of context. On Zurelay one answer runs to at most 32,000 output tokens, whatever max_tokens you set, and then ends with finish_reason "length". Reasoning and the answer share those 32,000 tokens, so leave headroom on hard prompts. For longer answers use kimi-k3.1-pro-highspeed, which goes past 32,000 tokens (Moonshot's own default is 131,072).

Can Kimi K3.1 Pro read images?

Yes. Kimi K3.1 Pro has native vision and takes images alongside text in the same request. Send them in OpenAI's format, as https links or base64 data URLs: Moonshot's API doesn't take links, so Zurelay downloads them and sends the image itself. The model returns text only.

Can I turn off Kimi K3.1 Pro's thinking?

No. It always thinks before it answers. Set reasoning_effort to low for the shortest thinking and the fastest, cheapest answers, or leave it at max, the default, for hard problems. On Zurelay, a request for no reasoning runs at low.

Does the Kimi K3.1 Pro API support streaming and per-key limits?

Yes. Set stream: true to receive tokens as they are generated, and add stream_options.include_usage to get token counts in the final chunk. You can give each Zurelay API key its own monthly budget and requests-per-minute cap. Failed requests are retried on another route and never billed.

Is Kimi K3.1 Pro on Zurelay the same model Moonshot serves?

Yes. Requests go to kimi-k3.1-pro itself. Zurelay doesn't swap in a smaller or substitute model. Only smart routing, which you turn on yourself, can answer with a close alternative when the model is down, and the x-zurelay-model header names the model that answered. Output is sampled, so wording varies from run to run, just as it does on Moonshot's own API. Zurelay is an independent service and is not affiliated with Moonshot AI.

More models on Zurelay

All models

Stop paying list price.
Start saving today.

Create a key in seconds, point the SDK you already use at Zurelay, and every request costs up to 90% less from the first token.

Get your API keyTalk to an engineer