Kimi K2.7-Code

Kimi K2.7-Code is a 1000B-parameter model you reach through an API, with 32B parameters active per token — too large to self-host, so this page covers what it does, how to access it, and what to run locally instead.

PARAMETERS
1000B (32B active)
FORMAT
API
RUNS LOCALLY
No
BEST FOR
Coding, Agentic tasks

You don't run Kimi K2.7-Code locally

At 1000B parameters (32B active per token), a Q4-class build of Kimi K2.7-Code would need roughly 600 GB — beyond any consumer machine (0.6 GB per billion parameters, our standard Q4 rule). Access is through the hosted API.

Jun 12, 2026 Moonshot AI release. 1T total / 32B active MoE built on K2.6. Moonshot reports it cuts thinking-token usage ~30%, a self-reported figure. 256K context, Modified MIT license. Open weights, but 1T scale means cloud/API for nearly all users; Ollama exposes a cloud-only tag.

Strong local alternatives

More Kimi models

Frequently asked questions

Can I run Kimi K2.7-Code locally?

Not realistically. Kimi K2.7-Code is a 1000B-parameter model (32B active); a Q4-class build would need roughly 600 GB — beyond any consumer machine. The hosted API or a smaller open model is the practical path.

How do I access Kimi K2.7-Code?

Through the vendor-hosted API. See the official source linked on this page.

What is the best local alternative to Kimi K2.7-Code?

Qwen3 235B A22B is the strongest local model we track (235B). It runs on a single high-memory Mac or GPU; see its page for exact hardware.

Cite this page

ModelFit: Kimi K2.7-Code — specs, memory math and hardware verdicts.
https://modelfit.io/models/kimi-k2.7-code/ (dataset updated 2026-09-03, CC BY 4.0).