Moonshot AI
Operational· checked every few minutes

Kimi K3.1 Pro Highspeed API

Kimi K3.1 Pro at two to three times the speed, 18% below Moonshot's list price

Price per 1M tokens18%off
Input
$3.30$4.00
Output
$16.50$20.00
Cached input
$0.165
Get your API keyTry it in the playground
Model IDkimi-k3.1-pro-highspeed
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-highspeed",
    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
Reasoning
low · high · max
Input
Text, Image
Released
Oct 24, 2026
Status
Operational

Pricing

Kimi K3.1 Pro Highspeed 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
$3.30$4.00
Output
per 1M tokens
$16.50$20.00
Cached input
per 1M tokens
$0.165—

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
$330.00
Moonshot AI list price
$400.00
You save every month$70.00

$840.00 a year

Overview

What is Kimi K3.1 Pro Highspeed?

Kimi K3.1 Pro Highspeed is the fast version of Moonshot AI's newest flagship: the same K3.1 Pro model, with the same 1M-token context and vision, served at two to three times the speed. The Zurelay API serves it through one OpenAI-compatible endpoint at 18% below Moonshot's list price.

Moonshot AI released Kimi K3.1 Pro Highspeed in October 2026 alongside Kimi K3.1 Pro, the first model of its K3.1 series. Moonshot says the two are equally capable and Highspeed writes two to three times faster. In our tests through Zurelay it wrote about 97 tokens a second against about 46 for the standard model, so a long reasoning answer arrives in about half the time.

The model is Kimi K3 post-trained with far more reinforcement learning. Moonshot reports 65.7% on Terminal-Bench 4.0, ahead of GPT-6 Astra and just behind Claude Opus 5.5, 92.7% on BenchCAD with tools, a state-of-the-art 64.4 on OfficeQA Pro v2, and a claim-level hallucination rate down from 23.4% on K3 to 8.3%. It is built for coding agents, games and web experiences, 3D and CAD modeling, and financial work.

Thinking is always on, with a reasoning_effort of low, high or max, and max as the default; low cuts the thinking to almost nothing for quick replies. It reads text and images, calls tools, follows a strict JSON schema in response_format, and holds 1,048,576 tokens of context. Temperature and top_p are fixed at 1.0 and 0.95. Moonshot doesn't allow a tool forced by name while thinking is on, so Zurelay sends that request as auto with the tool named; tool_choice required works as usual.

On Zurelay you call it at https://api.zurelay.com/v1 with the model ID kimi-k3.1-pro-highspeed, for $3.30 per 1M input tokens and $16.50 per 1M output tokens against Moonshot's list price of $4.00 and $20.00, with cached input at $0.165.

Strengths

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

01

Answers in half the time

About 97 tokens a second in our tests, twice the standard K3.1 Pro, which matters most on long reasoning answers.

02

The full K3.1 Pro

The same model as Kimi K3.1 Pro, not a smaller one: Moonshot says the two are equally capable.

03

Strict JSON

Follows a JSON schema in response_format, so the answer can go straight into your code.

04

1M-token context

Holds a whole repository, a long contract or a stack of filings in one request.

Use cases

  • Interactive coding assistants

    Keep a developer in flow: edits, explanations and fixes come back fast enough to use as you type.

  • Chat products

    Serve a strong reasoning model in a live chat without the long wait that thinking usually adds.

  • Agents with many steps

    Shorten every turn of a tool-calling loop, so a 50-step task finishes in about half the time.

  • Structured extraction

    Pull fields from documents or screenshots into your JSON schema while a user waits.

Get started

Call Kimi K3.1 Pro Highspeed 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-highspeed

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

    kimi-k3.1-pro-highspeed

Works with the tools you already use

Compare

Similar models. All on the same key.

60%off

Kimi K3.1 Pro

$0.80 / $4.00 per 1M tokens

Pick Kimi K3.1 Pro for the same model at a much lower price, when nobody is waiting on the answer and it fits in 32,000 tokens.

Kimi K3.1 Pro 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.

Kimi K3 API
60%off

Claude Haiku 5.5

$0.04 / $0.20 per 1M tokens

Pick Claude Haiku 5.5 for a fast, low-cost model when you don't need flagship reasoning.

Claude Haiku 5.5 API

FAQ

Kimi K3.1 Pro Highspeed API questions.

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

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

Kimi K3.1 Pro Highspeed costs $3.30 per 1M input tokens and $16.50 per 1M output tokens on Zurelay, with cached input at $0.165. Moonshot's list price is $4.00 input and $20.00 output, so you save 18%. One rate applies at every prompt length, up to the full 1M-token window.

How fast is Kimi K3.1 Pro Highspeed?

Moonshot says two to three times the speed of Kimi K3.1 Pro. In our tests through Zurelay it wrote about 97 tokens a second, against about 46 for K3.1 Pro, with the first tokens usually back in one to three seconds at low reasoning effort.

Is Highspeed a smaller or weaker model?

No. It is the same Kimi K3.1 Pro model, served faster, and Moonshot says the two are equally capable. You pay more for the speed: if nobody is waiting on the answer, kimi-k3.1-pro gives the same results for less.

How long can one Kimi K3.1 Pro Highspeed answer be?

Longer than Kimi K3.1 Pro's on Zurelay. Moonshot lets one answer run to 131,072 tokens by default, reasoning included, and in our tests a single Highspeed answer ran past 53,000 tokens in about five minutes. Set max_completion_tokens (or max_tokens) to the length you need. A kimi-k3.1-pro answer stops at 32,000 output tokens.

How do I call Kimi K3.1 Pro Highspeed 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-highspeed" to chat.completions.create. Leave temperature and top_p unset, since Moonshot fixes them, and use reasoning_effort to pick low, high or max.

Does Kimi K3.1 Pro Highspeed support JSON schema output?

Yes. Send response_format with type "json_schema" and your schema, and the final answer comes back as JSON that follows it. Parse the message content, not the reasoning.

Can I force a specific tool call?

tool_choice required works: the model calls at least one of your tools. Moonshot refuses a tool forced by name while thinking is on, so Zurelay sends that request as auto with the tool named, and the model almost always calls it.

Can Kimi K3.1 Pro Highspeed read images?

Yes. It takes images alongside text in the same request, 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. It returns text only.

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

Yes. Requests go to kimi-k3.1-pro-highspeed itself. Zurelay doesn't swap in a smaller or substitute model. If it's down and you've turned on smart routing, Zurelay can answer with Kimi K3.1 Pro, the same model at normal speed, and the x-zurelay-model header names the model that answered. 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