Kimi K2.5 - AI Model API and Pricing

Long source material in, readable article out

Its strengths are a long context and good summaries. Paste an interview transcript, meeting notes, a research brief or a webinar recording's text. Then ask for an article that keeps the key quotes and drops the filler. It is suited to the gap between raw material and a first publishable draft. That is where shorter-reach models start forgetting what was said early on.

Opening it in the AI Article Generator

Kimi K2.5 runs inside the AI Article Generator. Pick the Blog post preset for a narrative piece built from a transcript. Choose SEO article when you need headings around a search topic, or Newsletter for a digest of several sources. Product description, Ad copy and Video script are there too. Short, punchy copy is not this model's main strength, though.

A prompt that plays to its reach: "Below is a long interview transcript. Write an article for small shop owners covering the three practical tips the guest gives. Include one direct quote for each, and skip the small talk." Put the instruction above the pasted text so the task is clear before the long input.

Style, length and editing

The listed limit is a less familiar style. Phrasing can feel slightly off from what native readers expect, so plan an editing pass for voice. Giving it a short sample paragraph of your house style helps. There is no length setting to adjust. State a word count in the prompt to stop drafts running past what you will pay to edit.

What a post or a feature costs

An 800-word blog post, rounded up to 1k words, costs 6 credits, about $0.06 at the best pack. A 2,000-word article costs 12 credits, about $0.13. It is 4th from the cheapest of 10 text models billed per 1k words.

Other writers on the same balance

DeepSeek V4, at 2 credits, is cheaper and strong on reasoned long-form analysis. Its style can read as technical, though. Gemini 3 Flash at 5 credits is fast for high-volume descriptions and captions. Claude Haiku 4.5 at 8 credits writes natural sentences for captions, headlines and ad variants. It has a shorter reach than the larger models.

Where to actually run it

Kimi K2.5 API

curl -X POST https://ugesi.com/api/generate \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "kimi-k2-5",
    "prompt": "a product shot on white marble"
  }'

<?php
$ch = curl_init('https://ugesi.com/api/generate');
curl_setopt_array($ch, [
    CURLOPT_POST => true,
    CURLOPT_RETURNTRANSFER => true,
    CURLOPT_HTTPHEADER => [
        'Authorization: Bearer YOUR_API_KEY',
        'Content-Type: application/json',
    ],
    CURLOPT_POSTFIELDS => json_encode([
        'model' => 'kimi-k2-5',
        'prompt' => 'a product shot on white marble',
    ]),
]);
$response = json_decode(curl_exec($ch), true);
curl_close($ch);
print_r($response);

type GenerateRequest = {
  model: string;
  prompt: string;
};
type GenerateResponse = {
  job_id: number;
  job_token: string;
  credits_held: number;
  status: "queued" | "succeeded";
  output_text?: string;
};

const body: GenerateRequest = {
  model: "kimi-k2-5",
  prompt: "a product shot on white marble",
};

const resp = await fetch("https://ugesi.com/api/generate", {
  method: "POST",
  headers: {
    "Authorization": `Bearer ${process.env.UGESI_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify(body),
});
if (!resp.ok) throw new Error(`HTTP ${resp.status}`);
const job = (await resp.json()) as GenerateResponse;
console.log(job.job_id, job.status);

import requests

resp = requests.post(
    "https://ugesi.com/api/generate",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    json={
        "model": "kimi-k2-5",
        "prompt": "a product shot on white marble",
    },
    timeout=60,
)
resp.raise_for_status()
print(resp.json())

{
  "model": "kimi-k2-5",
  "prompt": "a product shot on white marble"
}

The API comes with every credit pack, from Lite to Enterprise, and uses the same balance as the website. API documentation and access →

License and terms of use

Kimi K2.5 is released under the Modified MIT License, which allows commercial use on the conditions set out in that license. The full rules are in our Terms of service.

People also look for