Velokey

Kimi K2.7 Code

MoonshotChatin 0.76 · out 3.2 Credits / 1M tokens≈$0.76 / $3.2Available
Get API keyAPI docs

Kimi K2.7 Code is Moonshot AI's open-source agentic model for long-horizon software engineering, with a 256K context and always-on thinking. Access it through Velokey's unified endpoint with below-list pricing.

Long-horizon coding1T-param MoE256K context~30% fewer thinking tokens

Use cases

Repository-scale refactoring · Multi-turn coding sessions

Input

Text, image

Output

Text

Billing

Per token

0.7
02
2048
2568192
kimi-k2.7-code

Start a conversation

Type a message below to begin

Sign up to get $0.5 in free credits — about 20 free images

Pricing details

Transparent usage-based pricing with no hidden fees.

Detailed Pricing

Usage-based pricing breakdown

SpecPriceOfficialSavings
Input/1M tokens0.76 Credits≈$0.760.95 Credits-20%
Output/1M tokens3.2 Credits≈$3.24 Credits-20%
Cache Read/1M tokens0.152 Credits≈$0.1520.19 Credits-20%

Billing Rules

  • Output price varies by model, resolution, quality, duration, or token usage.
  • Uploaded reference files may be charged separately when supported.
  • Add-on capabilities such as web search or image search may be billed per request.
  • Token-based models are billed by input and output tokens.
  • If the primary route is unavailable, Velokey may automatically switch to a stable fallback route when possible.
  • Usage is pay-as-you-go with transparent credit deduction.

* Final cost depends on the generated result.

Related models

Explore other models in the same family.

Affordable Kimi K2.7 Code API

Access Moonshot AI's Kimi K2.7 Code on Velokey — long-horizon agentic coding, a 1T-parameter MoE architecture, a 256K context, and always-on thinking, with a unified endpoint and below-list pricing for production apps.

Meet Kimi K2.7 Code and its API access

Kimi K2.7 Code is Moonshot AI's open-source, coding-focused agentic model built for long-horizon software engineering. It excels at sustained-attention workflows — codebase refactoring, multi-file feature implementation, and extended debugging sessions — staying coherent across long agentic runs. It runs on a 1-trillion-parameter Mixture-of-Experts architecture that activates 32 billion parameters per token, pairs a 256K context window with always-on thinking, and cuts thinking-token usage by roughly 30% versus K2.6. Velokey provides an affordable, easy-to-integrate Kimi K2.7 Code API: a unified chat endpoint, a clear request flow, and below-list pricing for production systems, agent pipelines, and automation tools.

Long-horizon software engineering

Built for sustained-attention workflows — codebase refactoring, multi-file feature implementation, and extended debugging sessions that stay coherent across long agentic coding runs.

View Documentation →
Long-horizon software engineering

1T-parameter MoE with 256K context

A Mixture-of-Experts design with 384 experts (8 selected per token) across 61 layers activates 32B of 1T parameters per token, paired with Multi-head Latent Attention and a 256K context window.

View Documentation →
1T-parameter MoE with 256K context

~30% fewer thinking tokens

Improved reasoning efficiency cuts thinking-token usage by roughly 30% versus K2.6, keeping long agentic coding runs both capable and cost-efficient.

View Documentation →
~30% fewer thinking tokens

Always-on thinking for agentic coding

The model always runs with thinking enabled and does not support a non-thinking mode, giving consistent reasoning for autonomous agents and multi-turn coding sessions.

View Documentation →
Always-on thinking for agentic coding

How to deploy the Kimi K2.7 Code API on Velokey

Get started with just a few steps.

1

Sign up and get an API Key

Sign in to the Velokey console to generate an API Key that authenticates every request sent to Kimi K2.7 Code.

2

Configure request parameters

Set model to kimi-k2-7-code and submit messages. The model always runs with thinking enabled and does not support a non-thinking mode; use the 256K context for long-horizon work.

3

Integrate and start calling

Send requests to Velokey's unified /v1/chat/completions endpoint to power coding agents and automation, with usage managed in one console.

Real-world use cases for Kimi K2.7 Code

From repository refactoring to autonomous agents, covering complex software-engineering scenarios.

Repository-scale refactoring

With a 256K context and long-horizon focus, fit for refactoring and multi-file feature work across an entire repository.

Multi-file feature implementation

Coherent across many files at once — fit for implementing features that span services, modules, and shared interfaces.

Autonomous coding agents

Higher agentic-task scores make it fit for autonomous agents that plan and execute long task chains with always-on thinking.

Extended debugging sessions

Sustained attention keeps long debugging sessions on track, tracing root causes across large codebases without losing context.

Long-context repository work

A 256K context lets it reason across whole repositories, long files, and multi-turn coding history in a single session.

Reasoning-efficient automation

~30% fewer thinking tokens than K2.6 make long agentic coding pipelines more cost-efficient at scale.

Why choose Velokey for Kimi K2.7 Code

Discount on list price

Call Kimi K2.7 Code at a lower price on Velokey — pay-as-you-go with no forced subscription, reducing upfront cost.

Unified API interface

One API Key accesses Kimi K2.7 Code and other models — no juggling multiple accounts and keys, simplifying integration.

Complete docs and migration guidance

Endpoint references, parameter details, and examples help you adopt the 256K context and always-on thinking smoothly.

Stable and highly available

Automatic failover and stable fallback routes maintain availability when the primary route is down, keeping production reliable.

API Reference

Complete API call examples and parameter descriptions

Endpoint

https://api.velokey.ai/v1/chat/completions

Authentication

Bearer YOUR_API_KEY

Request Example

curl https://api.velokey.ai/v1/chat/completions \
  -X POST \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "kimi-k2.7-code",
    "messages": [{"role": "user", "content": "Hello!"}],
    "temperature": 0.7
  }'

Response Example

{
  "id": "req_abc123",
  "model": "kimi-k2.7-code",
  "created": 1234567890,
  "data": { ... }
}

Kimi K2.7 Code API FAQ

What is Kimi K2.7 Code built for?

It is Moonshot AI's open-source, coding-focused agentic model built for long-horizon software engineering — sustained-attention workflows such as codebase refactoring, multi-file feature implementation, and extended debugging sessions.

What is the architecture?

A 1-trillion-parameter Mixture-of-Experts model activating 32B parameters per token, with 384 experts (8 selected per token) across 61 layers (including 1 dense), Multi-head Latent Attention (MLA), and a 400M-parameter MoonViT vision encoder.

How large is the context window?

The context window is 256K (262,144 tokens), fit for long-context repository work and multi-turn coding sessions.

How does it improve on K2.6?

It posts gains of +21.8% on Kimi Code Bench v2, +11.0% on Program Bench, and +31.5% on MLS Bench Lite, plus roughly 10% on agentic tasks (Kimi Claw, MCP Atlas, MCP Mark Verified), while cutting thinking-token usage by ~30% for better reasoning efficiency.

Does it support a non-thinking mode?

No. Kimi K2.7 Code always runs with thinking enabled and does not support a non-thinking mode, which suits agentic coding that needs consistent reasoning.

How is it billed on Velokey?

Official pricing per million tokens is $0.19 input on cache hit, $0.95 input on cache miss, and $4 output. This page's pricing area shows the current Velokey price, billed by input and output tokens with a discount on list price.

Start building with Kimi K2.7 Code today

Use one API key to access low-cost, stable AI models through Velokey.

Get API KeyView Docs