Moonshot/Dense

MKimi-K2 1026.5B

Name: Kimi-K2 1026.5B
Author: Moonshot

chat

1026.5B

Parameters

128K

Context length

Benchmarks

Quantizations

95K

HF downloads

Architecture

Dense

Released

2025-07-11

Layers

KV Heads

Head Dim

112

Family

kimi

Quantization Options

Quant	Bits	VRAM	Quality
IQ2_XXS	2.38	305.9 GB	low
IQ2_M	2.93	376.4 GB	low
Q2_K	3.16	406.0 GB	low
IQ3_XXS	3.25	417.5 GB	low
IQ3_XS	3.5	449.6 GB	low
Q3_K_S	3.64	467.5 GB	low
IQ3_M	3.76	482.9 GB	low
Q3_K_M	4	513.7 GB	low
Q3_K_L	4.3	552.2 GB	moderate
IQ4_XS	4.46	572.8 GB	moderate
Q4_K_S	4.67	599.7 GB	moderate
Q4_K_M	4.89	627.9 GB	good
Q5_K_S	5.57	715.2 GB	good
Q5_K_M	5.7	731.9 GB	good
Q6_K	6.56	842.2 GB	excellent
Q8_0	8.5	1091.1 GB	lossless
FP16	16	2053.5 GB	lossless

Select your GPU above to see speed estimates and compatibility for each quantization.

Benchmarks (8)

MATH-50097.4

IFEval89.8

MMLU-PRO81.1

GPQA Diamond75.1

SWE-bench65.8

LiveCodeBench53.7

AIME49.5

HLE4.7

Run this model

Easiest way to get starteddocs →

curl -fsSL https://ollama.com/install.sh | sh

$ollama run kimi:1026b-q4_k_m

Downloads and runs automatically. Add --verbose for speed stats.

Setup guide

HuggingFace Ollama Library GGUF Downloads Build Hardware

Find the best GPU for Kimi-K2 1026.5B

Build Hardware for Kimi-K2 1026.5B

MKimi-K2 1026.5B

Quantization Options

Benchmarks (8)

Run this model

Kimi-K2 1026.5B — 1026.5B Parameter Dense LLM

Model Specifications

VRAM Requirements

Benchmark Scores

How to Run Kimi-K2 1026.5B