Kimi model access
kimi-k2.7-code
Request kimi-k2.7-code through SteadyGateway’s production access process. Moonshot documents this exact model ID as a dedicated coding model with a 256K-token context window and OpenAI- and Anthropic-compatible APIs. Its weights are also published under a Modified MIT license with a large-scale attribution condition.
Source: Moonshot AI model catalog · verified 2026-08-09
What distinguishes kimi-k2.7-code?
These facts separate the coding model from Kimi K2.6. We track the thinking defaults because they change how production clients retain reasoning across tool-call turns; both sides of each comparison are sourced and dated.
| Fact | kimi-k2.7-code | Sibling contrast |
|---|---|---|
| Model role | Dedicated coding modelMoonshot AI model catalog · verified 2026-08-09 | kimi-k2.6: General multimodal and agentic modelMoonshot AI Kimi K2.6 model card · verified 2026-08-09 |
| Thinking mode | Always enabled; disabling it returns an errorMoonshot AI thinking-model documentation · verified 2026-08-09 | kimi-k2.6: Can run with thinking enabled or disabledMoonshot AI thinking-model documentation · verified 2026-08-09 |
| Preserved thinking | Always enabled; reasoning content must be retained across turnsMoonshot AI thinking-model documentation · verified 2026-08-09 | kimi-k2.6: Disabled by default and enabled with thinking.keepMoonshot AI thinking-model documentation · verified 2026-08-09 |
What are the facts behind kimi-k2.7-code?
| Fact | Value and source |
|---|---|
| Provider | Moonshot AI official APIMoonshot AI model catalog · verified 2026-08-09 |
| License | Modified MIT; commercial products or services with more than 100 million monthly active users or more than 20 million US dollars in monthly revenue must display “Kimi K2.7 Code” prominently in the user interfaceMoonshot AI Kimi K2.7 Code license · verified 2026-08-09 |
| Context window | 256K tokensMoonshot AI model catalog · verified 2026-08-09 |
| Modalities | Text, image, and video input; text outputMoonshot AI Kimi K2.7 Code model card · verified 2026-08-09 |
| Release | 2026-06-12Kimi official release announcement · verified 2026-08-09 |
| API interface | OpenAI- and Anthropic-compatible APIsMoonshot AI Kimi K2.7 Code model card · verified 2026-08-09 |
| Documented capabilities | Always-on thinking, preserved reasoning across turns, and multi-step tool callsMoonshot AI thinking-model documentation · verified 2026-08-09 |
How does kimi api access work for production teams?
If you searched for kimi api, the answer here is a production access conversation — not a free key or self-serve signup. SteadyGateway is for teams putting kimi-k2.7-code behind a production workload: request access with your model scope, region requirements and expected volume, and we return a written contract with the quota pool, version policy and support path defined before traffic moves.
OpenAI-compatible request shape · endpoint set by contract
const client = new OpenAI({
baseURL: process.env.STEADYGATEWAY_API_BASE_URL,
apiKey: process.env.STEADYGATEWAY_API_KEY,
})
await client.chat.completions.create({
model: 'kimi-k2.7-code',
messages: [{ role: 'user', content: '...' }],
})Pin the model ID when reproducibility matters. We handle the upstream capacity and quota design around that choice, so a fixed version is a procurement decision rather than a surprise operating limit.
Where is kimi-k2.7-code availability recorded?
The freshness board shows the release date, provider catalog state and verification date for this model. Open the row for the evidence behind the access conversation.
| Model | Released | Model Studio | AWS Bedrock | Azure AI Foundry |
|---|---|---|---|---|
| kimi-k2.7-code | 2026-06-12Kimi official release announcement · 2026-08-09 | unverified | unverified | availableMicrosoft Foundry model catalog · 2026-08-09 |
Every cell carries its source and verification date. Cells marked unverified are exactly that — we publish what we have checked, nothing more.
Open the kimi-k2.7-code ledger row → · Read the benchmark method → · See how production pricing is scoped →
What do buyers ask about kimi-k2.7-code?
How do I get kimi-k2.7-code API access for production?
Request production access. We scope the model, regions, quota pool and version policy to your workload, then return the commercial terms and credential path in writing.
Does kimi-k2.7-code support OpenAI- and Anthropic-compatible APIs?
Yes. Moonshot AI documents both interface families for Kimi K2.7 Code. The official OpenAI-compatible endpoint uses the exact model ID kimi-k2.7-code.
What is the kimi-k2.7-code context window?
Moonshot AI lists a 256K-token context window. The model accepts text, image, and video input and produces text output.
What does the kimi-k2.7-code Modified MIT license require?
If the software or a derivative is used in a commercial product or service with more than 100 million monthly active users, or more than 20 million US dollars in monthly revenue, the user interface must prominently display “Kimi K2.7 Code.”
How is kimi-k2.7-code different from kimi-k2.6?
Kimi K2.7 Code is the dedicated coding model. Its thinking and preserved-thinking modes are always on; Kimi K2.6 is a general multimodal agentic model that can disable thinking and leaves preserved thinking off by default.
Where are the kimi-k2.7-code weights and provider status documented?
Moonshot AI publishes the weights in its official Hugging Face repository. The freshness ledger records dated provider-catalog checks separately; as verified 2026-08-09, Microsoft Foundry lists the model while Alibaba Cloud Model Studio and AWS Bedrock remain unverified.
More Kimi coverage is in the model catalog.
Scope kimi-k2.7-code for production.
Tell us the workload, region needs and model policy. We will come back with a written access plan.
hello@steadygateway.com99.9% availability commitment with tiered service credits · Reply within one business day · NDA available on request