model family

Kimi K2

Published by moonshotai. No machine on this network has measured this family yet. The members below are what the catalog knows exists; nothing about what they can do is claimed here until somebody runs one.

what it takes, what it returns

text → text

Declared by the network’s catalog, not measured — a signature says what this family would be asked. Whether it answered is the signed capability flags on each member.

0 of 47 builds measured · 47 still open

Qwen3-32B-Kimi-K2-Thinking-Distill7 known builds · not measured

Kimi K2 BASE10 known builds · not measured

Kimi-K2-Instruct-FP41 known build · not measured

moonshotai.Kimi-K2-Instruct-BF161 known build · not measured

Kimi-K2-GPTQ-4b-128g-experts1 known build · not measured

kimi-k2-tiny-random1 known build · not measured

Kimi K2 INSTRUCT-090512 known builds · not measured

Kimi-K2-Instruct-0905-mlx-DQ3_K_M1 known build · not measured

Kimi-K2-Instruct-0905-MLX-5bit1 known build · not measured

Kimi K2 PRETRAINED1 known build · not measured

Kimi K2 THINKING10 known builds · not measured

Kimi-K2-Thinking-Eagle31 known build · not measured

Builds this network knows of

Declared, not measured — this is what the catalog says exists and what it would cost a machine to hold. Nothing here is signed, and none of it claims the build works.

how much memory does your machine have

4 of 47 builds fit a 16 GB machine — budgeting 70% of it, or 11.2 GiB, because the cache and the runtime want the rest.

yujiepan/kimi-k2-tiny-randomBF16vllmderived by yujiepan
0.0 GiB
nvidia/Kimi-K2-Thinking-Eagle3BF16vllmderived by nvidia
3.4 GiB
hf.co/mradermacher/Qwen3-32B-Kimi-K2-Thinking-Distill-i1-GGUF:IQ1_SIQ1_Sollamaderived by mradermacher
6.8 GiB
hf.co/mradermacher/Qwen3-32B-Kimi-K2-Thinking-Distill-i1-GGUF:IQ2_XXSIQ2_XXSollamaderived by mradermacher
8.4 GiB
hf.co/mradermacher/Qwen3-32B-Kimi-K2-Thinking-Distill-i1-GGUF:IQ3_XXSIQ3_XXSollamaderived by mradermacher
11.9 GiB
hf.co/mradermacher/Qwen3-32B-Kimi-K2-Thinking-Distill-i1-GGUF:Q4_0Q4_0ollamaderived by mradermacher
17.4 GiB
hf.co/mradermacher/Qwen3-32B-Kimi-K2-Thinking-Distill-i1-GGUF:Q4_K_MQ4_K_Mollamaderived by mradermacher
18.4 GiB
hf.co/TeichAI/Qwen3-32B-Kimi-K2-Thinking-Distill-GGUF:Q8_0Q8_0ollamaderived by TeichAI
32.4 GiB
hf.co/TeichAI/Qwen3-32B-Kimi-K2-Thinking-Distill-GGUF:BF16BF16ollamaderived by TeichAI
61.0 GiB
unsloth/Kimi-K2-Instruct-GGUF:IQ1_SIQ1_Sllamacppquantized by unsloth
260.9 GiB
unsloth/Kimi-K2-Instruct-0905-GGUF:IQ1_SIQ1_Sllamacppquantized by unsloth
262.5 GiB
unsloth/Kimi-K2-Thinking-GGUF:IQ1_SIQ1_Sllamacppquantized by unsloth
265.7 GiB
unsloth/Kimi-K2-Instruct-GGUF:IQ2_XXSIQ2_XXSllamacppquantized by unsloth
306.2 GiB
unsloth/Kimi-K2-Instruct-0905-GGUF:IQ2_XXSIQ2_XXSllamacppquantized by unsloth
308.1 GiB
unsloth/Kimi-K2-Thinking-GGUF:IQ2_XXSIQ2_XXSllamacppquantized by unsloth
312.2 GiB
unsloth/Kimi-K2-Instruct-GGUF:IQ3_XXSIQ3_XXSllamacppquantized by unsloth
388.0 GiB
unsloth/Kimi-K2-Instruct-0905-GGUF:IQ3_XXSIQ3_XXSllamacppquantized by unsloth
388.6 GiB
unsloth/Kimi-K2-Thinking-GGUF:IQ3_XXSIQ3_XXSllamacppquantized by unsloth
392.7 GiB
mlx-community/moonshotai_Kimi-K2-Instruct-mlx-3bit3BITmlxquantized by mlx-community
418.5 GiB
mlx-community/Kimi-K2-Instruct-0905-mlx-DQ3_K_M4BITmlxderived by mlx-community
436.2 GiB
UnicornChan/Kimi-K2-GPTQ-4b-128g-expertsW4A16vllmderived by UnicornChan
509.1 GiB
amd/Kimi-K2-Instruct-0905-MXFP4QUARKvllmquantized by amd
520.3 GiB
mlx-community/Kimi-K2-Instruct-4bit4BITmlxquantized by mlx-community
537.9 GiB
richardyoung/Kimi-K2-Instruct-0905-MLX-4bit4BITmlxquantized by richardyoung
537.9 GiB
unsloth/Kimi-K2-Instruct-GGUF:Q4_0Q4_0llamacppquantized by unsloth
540.7 GiB
unsloth/Kimi-K2-Instruct-0905-GGUF:Q4_0Q4_0llamacppquantized by unsloth
541.3 GiB
unsloth/Kimi-K2-Thinking-GGUF:Q4_0Q4_0llamacppquantized by unsloth
541.3 GiB
baseten/Kimi-K2-Instruct-FP4FP8vllmderived by baseten
549.4 GiB
moonshotai/Kimi-K2-ThinkingW4A16vllmpublished by moonshotai
553.4 GiB
nvidia/Kimi-K2-Thinking-NVFP4NVFP4vllmquantized by nvidia
553.4 GiB
unsloth/Kimi-K2-Instruct-GGUF:Q4_K_MQ4_K_Mllamacppquantized by unsloth
578.1 GiB
unsloth/Kimi-K2-Instruct-0905-GGUF:Q4_K_MQ4_K_Mllamacppquantized by unsloth
578.6 GiB
unsloth/Kimi-K2-Thinking-GGUF:Q4_K_MQ4_K_Mllamacppquantized by unsloth
578.6 GiB
mlx-community/Kimi-K2-Thinking4BITmlxquantized by mlx-community
612.5 GiB
richardyoung/Kimi-K2-Instruct-0905-MLX-5bit5BITmlxderived by richardyoung
657.4 GiB
richardyoung/Kimi-K2-Instruct-0905-MLX-6bit6BITmlxquantized by richardyoung
776.9 GiB
moonshotai/Kimi-K2-InstructFP8vllmpublished by moonshotai
958.5 GiB
moonshotai/Kimi-K2-Instruct-0905FP8vllmpublished by moonshotai
958.5 GiB
moonshotai/Kimi-K2-BaseFP8vllmpublished by moonshotai
958.5 GiB
richardyoung/Kimi-K2-Instruct-0905-MLX-8bit8BITmlxquantized by richardyoung
1015.8 GiB
unsloth/Kimi-K2-Instruct-GGUF:Q8_0Q8_0llamacppquantized by unsloth
1016.1 GiB
unsloth/Kimi-K2-Instruct-0905-GGUF:Q8_0Q8_0llamacppquantized by unsloth
1016.1 GiB
unsloth/Kimi-K2-Thinking-GGUF:Q8_0Q8_0llamacppquantized by unsloth
1016.1 GiB
DevQuasar/moonshotai.Kimi-K2-Instruct-BF16BF16vllmderived by DevQuasar
1911.8 GiB
unsloth/Kimi-K2-Instruct-GGUF:BF16BF16llamacppquantized by unsloth
1912.1 GiB
unsloth/Kimi-K2-Instruct-0905-GGUF:BF16BF16llamacppquantized by unsloth
1912.1 GiB
unsloth/Kimi-K2-Thinking-GGUF:BF16BF16llamacppquantized by unsloth
1912.1 GiB

Sizes are what the catalog declares, not what a machine measured — fitting on disk is necessary and not sufficient. Nobody here has run these on your machine, and this narrows the candidates rather than certifying one.

the same builds, by member

Qwen3-32B-Kimi-K2-Thinking-Distill

hf.co/mradermacher/Qwen3-32B-Kimi-K2-Thinking-Distill-i1-GGUF:IQ1_SIQ1_Sollama6.8 GiB on diskderived by mradermacher

hf.co/mradermacher/Qwen3-32B-Kimi-K2-Thinking-Distill-i1-GGUF:IQ2_XXSIQ2_XXSollama8.4 GiB on diskderived by mradermacher

hf.co/mradermacher/Qwen3-32B-Kimi-K2-Thinking-Distill-i1-GGUF:IQ3_XXSIQ3_XXSollama11.9 GiB on diskderived by mradermacher

hf.co/mradermacher/Qwen3-32B-Kimi-K2-Thinking-Distill-i1-GGUF:Q4_0Q4_0ollama17.4 GiB on diskderived by mradermacher

hf.co/mradermacher/Qwen3-32B-Kimi-K2-Thinking-Distill-i1-GGUF:Q4_K_MQ4_K_Mollama18.4 GiB on diskderived by mradermacher

hf.co/TeichAI/Qwen3-32B-Kimi-K2-Thinking-Distill-GGUF:Q8_0Q8_0ollama32.4 GiB on diskderived by TeichAI

hf.co/TeichAI/Qwen3-32B-Kimi-K2-Thinking-Distill-GGUF:BF16BF16ollama61.0 GiB on diskderived by TeichAI

Kimi K2 BASE

unsloth/Kimi-K2-Instruct-GGUF:IQ1_SIQ1_Sllamacpp260.9 GiB on diskquantized by unsloth

unsloth/Kimi-K2-Instruct-GGUF:IQ2_XXSIQ2_XXSllamacpp306.2 GiB on diskquantized by unsloth

unsloth/Kimi-K2-Instruct-GGUF:IQ3_XXSIQ3_XXSllamacpp388.0 GiB on diskquantized by unsloth

mlx-community/moonshotai_Kimi-K2-Instruct-mlx-3bit3BITmlx418.5 GiB on diskquantized by mlx-community

mlx-community/Kimi-K2-Instruct-4bit4BITmlx537.9 GiB on diskquantized by mlx-community

unsloth/Kimi-K2-Instruct-GGUF:Q4_0Q4_0llamacpp540.7 GiB on diskquantized by unsloth

unsloth/Kimi-K2-Instruct-GGUF:Q4_K_MQ4_K_Mllamacpp578.1 GiB on diskquantized by unsloth

moonshotai/Kimi-K2-InstructFP8vllm958.5 GiB on diskpublished by moonshotai

unsloth/Kimi-K2-Instruct-GGUF:Q8_0Q8_0llamacpp1016.1 GiB on diskquantized by unsloth

unsloth/Kimi-K2-Instruct-GGUF:BF16BF16llamacpp1912.1 GiB on diskquantized by unsloth

Kimi-K2-Instruct-FP4

baseten/Kimi-K2-Instruct-FP4FP8vllm549.4 GiB on diskderived by baseten

moonshotai.Kimi-K2-Instruct-BF16

DevQuasar/moonshotai.Kimi-K2-Instruct-BF16BF16vllm1911.8 GiB on diskderived by DevQuasar

Kimi-K2-GPTQ-4b-128g-experts

UnicornChan/Kimi-K2-GPTQ-4b-128g-expertsW4A16vllm509.1 GiB on diskderived by UnicornChan

kimi-k2-tiny-random

yujiepan/kimi-k2-tiny-randomBF16vllm20 MiB on diskderived by yujiepan

Kimi K2 INSTRUCT-0905

unsloth/Kimi-K2-Instruct-0905-GGUF:IQ1_SIQ1_Sllamacpp262.5 GiB on diskquantized by unsloth

unsloth/Kimi-K2-Instruct-0905-GGUF:IQ2_XXSIQ2_XXSllamacpp308.1 GiB on diskquantized by unsloth

unsloth/Kimi-K2-Instruct-0905-GGUF:IQ3_XXSIQ3_XXSllamacpp388.6 GiB on diskquantized by unsloth

amd/Kimi-K2-Instruct-0905-MXFP4QUARKvllm520.3 GiB on diskquantized by amd

richardyoung/Kimi-K2-Instruct-0905-MLX-4bit4BITmlx537.9 GiB on diskquantized by richardyoung

unsloth/Kimi-K2-Instruct-0905-GGUF:Q4_0Q4_0llamacpp541.3 GiB on diskquantized by unsloth

unsloth/Kimi-K2-Instruct-0905-GGUF:Q4_K_MQ4_K_Mllamacpp578.6 GiB on diskquantized by unsloth

richardyoung/Kimi-K2-Instruct-0905-MLX-6bit6BITmlx776.9 GiB on diskquantized by richardyoung

moonshotai/Kimi-K2-Instruct-0905FP8vllm958.5 GiB on diskpublished by moonshotai

richardyoung/Kimi-K2-Instruct-0905-MLX-8bit8BITmlx1015.8 GiB on diskquantized by richardyoung

unsloth/Kimi-K2-Instruct-0905-GGUF:Q8_0Q8_0llamacpp1016.1 GiB on diskquantized by unsloth

unsloth/Kimi-K2-Instruct-0905-GGUF:BF16BF16llamacpp1912.1 GiB on diskquantized by unsloth

Kimi-K2-Instruct-0905-mlx-DQ3_K_M

mlx-community/Kimi-K2-Instruct-0905-mlx-DQ3_K_M4BITmlx436.2 GiB on diskderived by mlx-community

Kimi-K2-Instruct-0905-MLX-5bit

richardyoung/Kimi-K2-Instruct-0905-MLX-5bit5BITmlx657.4 GiB on diskderived by richardyoung

Kimi K2 PRETRAINED

moonshotai/Kimi-K2-BaseFP8vllm958.5 GiB on diskpublished by moonshotai

Kimi K2 THINKING

unsloth/Kimi-K2-Thinking-GGUF:IQ1_SIQ1_Sllamacpp265.7 GiB on diskquantized by unsloth

unsloth/Kimi-K2-Thinking-GGUF:IQ2_XXSIQ2_XXSllamacpp312.2 GiB on diskquantized by unsloth

unsloth/Kimi-K2-Thinking-GGUF:IQ3_XXSIQ3_XXSllamacpp392.7 GiB on diskquantized by unsloth

unsloth/Kimi-K2-Thinking-GGUF:Q4_0Q4_0llamacpp541.3 GiB on diskquantized by unsloth

moonshotai/Kimi-K2-ThinkingW4A16vllm553.4 GiB on diskpublished by moonshotai

nvidia/Kimi-K2-Thinking-NVFP4NVFP4vllm553.4 GiB on diskquantized by nvidia

unsloth/Kimi-K2-Thinking-GGUF:Q4_K_MQ4_K_Mllamacpp578.6 GiB on diskquantized by unsloth

mlx-community/Kimi-K2-Thinking4BITmlx612.5 GiB on diskquantized by mlx-community

unsloth/Kimi-K2-Thinking-GGUF:Q8_0Q8_0llamacpp1016.1 GiB on diskquantized by unsloth

unsloth/Kimi-K2-Thinking-GGUF:BF16BF16llamacpp1912.1 GiB on diskquantized by unsloth

Kimi-K2-Thinking-Eagle3

nvidia/Kimi-K2-Thinking-Eagle3BF16vllm3.4 GiB on diskderived by nvidia

ask your own gram

Your concierge can answer this from the network rather than from this page. The ask carries the operator id, so it resolves to a capability and an engine — not to a name somebody has to recognise.

What has this network measured about the Kimi K2 family (epn.family.kimi-k2), and can my gram run any of it? See /inferences/kimi-k2

open a gramx with this ask →epn.family.kimi-k2