Kimi K3 TL0

TL0 Open weights

Open-weight multimodal reasoning and long-horizon coding with 1M-token context

Pricing

Prices for this trust level. Other trust levels are listed below. Prices are per 1M tokens.

Trust levels for this model
Trust Level Region Context Max outputInput / 1MCache / 1MOutput / 1M§203
TL0 US — — 2.00 € 2.20 $ 0.25 € 0.28 $ 12.00 € 13.22 $ —

USD prices are converted from EUR at the ECB reference rate of 1 EUR = 1.1014 USD (2026-11-16). EUR prices are authoritative.

Prices as of 16.11.2026.

Other Trust Levels

Prices for the other trust levels available for this model.

Other trust levels for this model
Trust Level Region Context Max outputInput / 1MCache / 1MOutput / 1M§203
TL1 SG — — 3.00 € 3.30 $ 0.30 € 0.33 $ 15.00 € 16.52 $ —

About this model

Kimi K3 is Moonshot AI's 2.8T-parameter open-weight multimodal model for long-horizon software engineering, reasoning, and agentic knowledge work. Its sparse Mixture-of-Experts architecture activates 104B parameters per token, selecting 16 of 896 experts. Kimi Delta Attention, Attention Residuals, and Stable LatentMoE enable efficient scaling and a 1M-token context window. The model processes text and images and generates text, supporting large-repository coding, debugging, terminal-based workflows, and vision-guided application development. It also handles document analysis, deep research, dashboards, and other multi-step tasks through tool calling. Thinking is always enabled, with low, high, and max reasoning-effort settings. Quantization-aware training uses MXFP4 weights and MXFP8 activations. The weights are released under the Kimi K3 License.

API

Model ID kimi-k3-tl0 OpenAI compatibleAnthropic compatible

Base URL https://api.aki.io/v1

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_AKI_API_KEY",
    base_url="https://api.aki.io/v1",
)

response = client.chat.completions.create(
    model="kimi-k3-tl0",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(response.choices[0].message.content)
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "YOUR_AKI_API_KEY",
  baseURL: "https://api.aki.io/v1",
});

const response = await client.chat.completions.create({
  model: "kimi-k3-tl0",
  messages: [{ role: "user", content: "Hello!" }],
});
console.log(response.choices[0].message.content);
curl https://api.aki.io/v1/chat/completions \
  -H "Authorization: Bearer $AKI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "kimi-k3-tl0",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Frequently Asked Questions

How much does Kimi K3 TL0 cost?
Input: 2.00 € per 1M tokens, Cache read: 0.25 € per 1M tokens, Output: 12.00 € per 1M tokens. Prices are exclusive of VAT.
What is the context window of Kimi K3 TL0?
Kimi K3 TL0 supports up to — tokens.
Which trust levels are available for Kimi K3 TL0?
Kimi K3 TL0 is available at TL1 (International Standard, SG), TL0 (Global Unrestricted, US).
Is Kimi K3 TL0 open weights?
Yes. The model weights are available on Hugging Face under the linked license.
Is Kimi K3 TL0 available with zero data retention?
Yes. Kimi K3 TL0 is served at TL0 (Global Unrestricted) with zero data retention and no model training.
How do I use Kimi K3 TL0 with the OpenAI SDK?
Use the model ID kimi-k3-tl0 with the OpenAI-compatible endpoint at https://api.aki.io/v1.

Start Building with Kimi K3 TL0