Kimi K2.7 Code
moonshotaiothertextvision

Kimi K2.7 Code

Updated Jun 18, 2026
Parameters
1058.6B
Context
262,144
License
other
Updated
Jun 18, 2026

Intelligence benchmarks

Artificial Analysis indexes - compared with the best open and proprietary models

Intelligence

41.9

AA Index

Coding

45.6

AA Index

Agentic

61.9

AA Index

Intelligence Index - Kimi K2.7 Code vs. the field

Best open-weight models (you can run locally) and leading proprietary models for context.

Claude Fable 5 (with fallback)
59.9
closed
Claude Opus 4.8 (max)
55.7
closed
GPT-5.5 (xhigh)
54.8
closed
Claude Opus 4.7 (max)
53.5
closed
GLM-5.2 (max)
50.7
open
Gemini 3.5 Flash
50.2
closed
MiniMax-M3
44.4
open
Kimi K2.7 Code
41.9
open

Coding Index comparison

Claude Fable 5 (with fallback)
76.5
closed
GPT-5.5 (xhigh)
74.9
closed
Gemini 3.1 Pro Preview
68.8
closed
GLM-5.2 (max)
67
open
Claude Opus 4.8 (max)
56.7
closed
GPT-5.3 Codex (xhigh)
53.1
closed
DeepSeek V4 Pro (Max)
47.5
open
Kimi K2.7 Code
45.6
open

Agentic Index comparison

Claude Opus 4.8 (max)
77.8
closed
GPT-5.5 (high)
72
closed
Claude Opus 4.7 (max)
71.3
closed
Gemini 3.5 Flash (medium)
70.4
closed
MiniMax-M3
68.6
open
MiMo-V2.5-Pro
67.4
open
DeepSeek V4 Pro (Max)
67.2
open
Kimi K2.7 Code
61.9
open

Benchmark data from Artificial Analysis · updated 2026-06-18.

Standard benchmarks

Performance across standard evaluations

BenchmarkScore
GPQA89.6

Will it run on your hardware?

Pick your GPU memory - see which quantizations fit, and the cheapest card for the rest

Too big for 24 GB at any quant
0 of 4 quantizations fit Kimi K2.7 Code with real runtime overhead.
Q4_K_M
614 GB
Q5_K_M
752 GB
Q8_0
1133 GB
FP16
2117 GB
fits tight too big

Need an exact figure for your context length? Use the VRAM calculator.

Run it locally

Copy-paste - running in under a minute

vLLMOpenAI-compatible API
vllm serve moonshotai/Kimi-K2.7-Code

New to this? Start with Ollama · serve to many users with vLLM.

Deep dive

Notes, sources, and the full write-up

Kimi K2.7 Code is a 1058.6B-parameter other model from Kimi. It scores 41.9 on the Artificial Analysis Intelligence Index (coding 45.6). At Q4_K_M it needs roughly 614 GB of VRAM, placing it in the 48 GB+ / multi-GPU hardware tier.

Benchmarks

Artificial Analysis Intelligence Index - Kimi K2.7 Code vs. leading closed models:

ModelIntelligenceCodingGPQA
Kimi K2.7 Code41.945.689.6
Claude Fable 5 (with fallback)59.976.592.6
Claude Opus 4.8 (max)55.756.792
GPT-5.5 (xhigh)54.874.993.5
Claude Opus 4.7 (max)53.552.591.4
Gemini 3.5 Flash50.24592.2

Source: Artificial Analysis (2026-06-18).

Popularity

Kimi K2.7 Code has 229,156 downloads in the last month on HuggingFace and 859 likes.

Frequently asked

Quick answers to common questions

How much VRAM does Kimi K2.7 Code need?

Kimi K2.7 Code with 1058.6B parameters needs approximately 614 GB at Q4_K_M quantization. Use our VRAM calculator for an exact estimate.

Is Kimi K2.7 Code better than other moonshotai models?

Kimi K2.7 Code has 1058.6B parameters with 262,144 context - a strong choice for general use.

What license is Kimi K2.7 Code under?

Kimi K2.7 Code is released under the other license, making it suitable for most commercial and personal projects.

What hardware runs Kimi K2.7 Code well?

With 1058.6B parameters, Kimi K2.7 Code requires adequate VRAM. High-end GPUs like the RTX 4090 (24GB), RTX 5090 (32GB), or Mac Studio with unified memory are good options. Check our hardware directory for specific recommendations.

What is the best quantization for Kimi K2.7 Code?

Q4_K_M is the recommended sweet spot - ~98% of FP16 quality at ~27% of the size. Q5_K_M (~752 GB) is an option if you have spare VRAM. Use our VRAM calculator to compare.

How long can Kimi K2.7 Code's context window handle?

Kimi K2.7 Code supports a 262,144-token context window - enough for very long documents, codebases, or multi-turn conversations. Real-world usable context may vary by implementation.

What models compete with Kimi K2.7 Code?

Kimi K2.7 Code competes with other 529B–1588B. Browse our model directory for comparisons, benchmarks, and community reviews to find the best fit.

Nearby options

Similar models and compatible hardware by spec

Comments coming soon

Configure NEXT_PUBLIC_GISCUS_REPO_ID and NEXT_PUBLIC_GISCUS_CATEGORY_ID at giscus.app to enable.