Moonshot AI: Kimi K3
moonshotai/kimi-k3
Access Kimi K3 from Moonshot AI using Puter.js AI API.
Get Started// npm install @heyputer/puter.js
import { puter } from '@heyputer/puter.js';
puter.ai.chat("Explain quantum computing in simple terms", {
model: "moonshotai/kimi-k3"
}).then(response => {
document.body.innerHTML = response.message.content;
});
<html>
<body>
<script src="https://js.puter.com/v2/"></script>
<script>
puter.ai.chat("Explain quantum computing in simple terms", {
model: "moonshotai/kimi-k3"
}).then(response => {
document.body.innerHTML = response.message.content;
});
</script>
</body>
</html>
# pip install openai
from openai import OpenAI
client = OpenAI(
base_url="https://api.puter.com/puterai/openai/v1/",
api_key="YOUR_PUTER_AUTH_TOKEN",
)
response = client.chat.completions.create(
model="moonshotai/kimi-k3",
messages=[
{"role": "user", "content": "Explain quantum computing in simple terms"}
],
)
print(response.choices[0].message.content)
curl https://api.puter.com/puterai/openai/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer YOUR_PUTER_AUTH_TOKEN" \
-d '{
"model": "moonshotai/kimi-k3",
"messages": [
{"role": "user", "content": "Explain quantum computing in simple terms"}
]
}'
Model Card
Kimi K3 is Moonshot AI's flagship open-weight model, released July 16, 2026, with full weights following on July 27. At roughly 2.8 trillion parameters in a Mixture-of-Experts architecture, Moonshot positions it as the largest open-source model released to date, built on two new components: Kimi Delta Attention, a hybrid linear attention mechanism, and Attention Residuals, a replacement for standard residual connections.
It runs in an always-on thinking mode with a 1-million-token context window and accepts text, image, and video input. Reported results include 93.5% on GPQA Diamond, 91.2% on BrowseComp, 88.3% on Terminal-Bench 2.1, and a first-place finish on Arena.ai's Frontend Code Arena, putting it close to Claude Opus 4.8 and GPT-5.5 on several agentic and coding tasks. These figures come from Moonshot and early testers, not independently confirmed leaderboards.
It suits developers building long-horizon coding agents and tool-calling pipelines who want frontier-level performance at open-weight pricing.
Context Window 1M
tokens
Max Output 1M
tokens
Input Cost $3
per million tokens
Output Cost $15
per million tokens
Input text, image, video
modalities
Tool Use Yes
Release Date Jul 16, 2026
Output Speed 59
tokens / sec
Latency 1.35s
time to first token
Model Playground
Try Kimi K3 instantly in your browser.
This playground uses the Puter.js AI API — no API keys or setup required.
Benchmarks
How Kimi K3 performs on standard evaluations.
| Benchmark | Score |
|---|---|
| GPQA Diamond Graduate-level science Q&A | 93.5% |
| Humanity's Last Exam Cross-domain reasoning | 44.3% |
| SciCode Scientific programming | 58.7% |
| LCR Long-context reasoning | 74.7% |
Scores sourced from Artificial Analysis.
Find other Moonshot AI models →
Kimi K2.7 Code
Kimi K2.7 Code is Moonshot AI's open-weight coding-agent model, released June 2026 and purpose-built for long-horizon, autonomous coding tasks. It shares the same 1-trillion-parameter Mixture-of-Experts architecture (32B active parameters) as K2.6 but is entirely focused on software engineering workloads. Compared to K2.6, it improves 21.8% on Kimi Code Bench v2, 11% on Program Bench, and 31.5% on MLS Bench Lite, while cutting reasoning-token usage by roughly 30%. It always runs in thinking mode — non-thinking mode is not supported. With a 262K-token context window, K2.7 Code is well-suited for multi-file, repository-scale coding pipelines and agentic workflows where sustained reasoning and deep code understanding matter.
ChatKimi K2.6
Kimi K2.6 is Moonshot AI's latest open-weight multimodal model, built on a 1-trillion-parameter mixture-of-experts architecture with a 256K context window. It excels at agentic coding and long-horizon execution, supporting sustained autonomous workflows with 4,000+ tool calls across languages like Rust, Go, and Python. On key benchmarks, it scores 58.6 on SWE-Bench Pro, 54.0 on HLE with Tools, and 50.0 on Toolathlon — competitive with GPT-5.4 and Claude Opus 4.6 on coding and agent tasks, though trailing them on pure reasoning. The model accepts text, image, and video input, supports both thinking and non-thinking modes, and offers an OpenAI-compatible API. It's a strong pick for developers building multi-step agentic workflows and complex software engineering pipelines.
ChatKimi K2.5
Kimi K2.5 is Moonshot AI's most capable open-source model, a natively multimodal (vision + text) trillion-parameter MoE with 32B active parameters released in January 2026. Built through continual pretraining on ~15 trillion mixed visual and text tokens atop the K2 base, it supports both thinking and instant modes with a 256K context window. It scored 76.8% on SWE-bench Verified, 96.1% on AIME 2025, and 50.2% on Humanity's Last Exam with tools — outperforming Claude Opus 4.5 and GPT-5.2 on the latter. Its standout feature is Agent Swarm, which coordinates up to 100 parallel sub-agents for complex tasks. K2.5 excels at vision-to-code generation, frontend development from screenshots, and large-scale agentic workflows, making it a strong choice for developers building multimodal AI agents.
Frequently Asked Questions
You can access Kimi K3 by Moonshot AI through Puter.js AI API. Include the library in your web app or Node.js project and start making calls with just a few lines of JavaScript — no backend and no configuration required. You can also use it with Python or cURL via Puter's OpenAI-compatible API.
Yes, it is free if you're using it through Puter.js. With the User-Pays Model, you can add Kimi K3 to your app at no cost — your users pay for their own AI usage directly, making it completely free for you as a developer.
| Price per 1M tokens | |
|---|---|
| Input | $3 |
| Output | $15 |
Kimi K3 was created by Moonshot AI and released on Jul 16, 2026.
Kimi K3 supports a context window of 1M tokens. For reference, that is roughly equivalent to 2,097 pages of text.
Kimi K3 can generate up to 1M tokens in a single response.
Kimi K3 accepts the following input types: text, image, video. It produces: text.
Yes, Kimi K3 supports tool use (function calling), allowing it to interact with external tools, APIs, and data sources as part of its response flow.
Kimi K3 scores 57.1 on the Artificial Analysis Intelligence Index, outperforming 99% of tracked models. On coding, it scores 76.2 (outperforms 96% of models).
Yes — the Kimi K3 API works with any JavaScript framework, Node.js, or plain HTML through Puter.js. Just include the library and start building. See the documentation for more details.
Get started with Puter.js
Add Kimi K3 to your app without worrying about API keys or setup.
Read the Docs View Tutorials