Moonshot AI: Kimi K2 Thinking

Thinking Kimi model for slower research passes, planning, and hard technical questions

Model lab: Moonshot AI
Model family: kimi-thinking

Model resources

Hugging Face

Modalitiestext → text

Context window262,144

Maximum output262,144

InputUSD 0.48 / 1M tokens

OutputUSD 2 / 1M tokens

Try now View docs

Quickstart

Use your Moduck API Key, set the API base URL to https://api.moduck.ai/v1, and set the model parameter to moonshotai/kimi-k2-thinking.

import os
import requests

base_url = "https://api.moduck.ai/v1"
api_key = os.environ["MODUCK_API_KEY"]
model_id = "moonshotai/kimi-k2-thinking"

response = requests.post(
    f"{base_url}/chat/completions",
    headers={
        "Authorization": f"Bearer {api_key}",
        "Content-Type": "application/json",
    },
    json={
        "model": model_id,
        "messages": [{"role": "user", "content": "Hello!"}],
        "stream": False,
    },
    timeout=60,
)
response.raise_for_status()
print(response.json())

Use the familiar OpenAI-style chat request format to make model integration easier.

Kimi K2 Thinking API Pricing

Type	Price information
Input	USD 0.48 / 1M tokens
Output	USD 2 / 1M tokens
Cached input	USD 0.048 / 1M tokens

Benchmarks

Intelligence Index · v4.322

Coding Index—

Agentic Index—

Benchmark	Score	Measurement	Source
SWE-Bench Verified	71.3	resolved

Performance

Output speed143.55 tokens/s

Time to first token1.38 s

Time to first answer token—

End-to-end response time—

Performance trends

Throughput143.55 tok/sCurrent

Time to first token1.38 sCurrent

End-to-end latency7.1 sCurrent

Throughput

Current143.55 tok/s

Window average139.4 tok/s

Time to first token

Current1.38 s

Window average1.3 s

End-to-end latency

Current7.1 s

Window average7.5 s

Tail latency (P99)

Current15.0 s

Window average17.2 s

Prompt cache hit rate

Current26.1%

Window average23.3%

Error rate

Current0.0%

Window average0.1%

Availability

Uptime100%At least one provider responded

Availability100%Requests actually served

Error rate0%Share of failed requests

Served requests, hour by hour

24 hours agoNow

Kimi K2 Thinking Compare models

Model	Input price	Output price	Context window	Capabilities
Kimi K2 ThinkingThis page	USD 0.48 / 1M tokens	USD 2 / 1M tokens	262,144	Text output · Reasoning · Tool calling
Kimi K2 Thinking Turbo	Price under review	Price under review	262,144	Text output · Reasoning · Tool calling
Kimi K2 Instruct 0905	USD 0.48 / 1M tokens	USD 2 / 1M tokens	262,144	Text output · Tool calling
Kimi K2 0711	USD 0.456 / 1M tokens	USD 1.84 / 1M tokens	131,072	Text output · Tool calling
Kimi K2 Instruct	USD 0.4 / 1M tokens	USD 1.6 / 1M tokens	131,072	Text output · Tool calling

Model specifications

Input: text
Output: text
Context window: 262,144
Maximum output: 262,144
Release date: 2025-11-06
Attachments: No
Reasoning: Yes
Tool calling: Yes
Temperature setting: Yes
Open weights: Yes

Frequently asked questions

What is Kimi K2 Thinking for?

Thinking Kimi model for slower research passes, planning, and hard technical questions

What inputs and outputs does Kimi K2 Thinking support?

Recorded inputs: text; outputs: text.

How is Kimi K2 Thinking priced?

The pricing table shows the current recorded Moduck price. For example: Input USD 0.48 / 1M tokens; Output USD 2 / 1M tokens. See the pricing table for all units, variants, and conditions.

What specifications are recorded for Kimi K2 Thinking?

Current model records list Context window 262,144 tokens; Maximum output 262,144 tokens.

What model parameter does the Quickstart use?

The example uses moonshotai/kimi-k2-thinking. This value comes from the page identifier and is not confirmed as a callable Moduck public ID.

Can I call this model through the Moduck API now?

This page has no verified Moduck API call for this model. A Quickstart example or reserved ID does not establish access; check the models available to your account.