Agents / Moonshot AI

Kimi K3

An open-weight multimodal model for long-running coding, reasoning, and knowledge work.

Verified sourceRepository opened Jun 13, 2026Source checked 9/23/2026

LOCALRENTED GPUOPEN WEIGHTS

Freshness 57 out of 100. Repository age: 148 days old, Jun 13, 2026. Measures age, not quality.

Popularity 84 out of 100. 771,618 downloads in 30 days; 11,940 lifetime likes. Checked Nov 7, 2026. Measures adoption, not quality.

Arena Open Source Agents: rank 5 of 23. Ranked model: Kimi K3 (Max). Export dated Nov 4, 2026. Source checked Nov 7, 2026. Human preference ranking, not hardware suitability.

Category rankings & sources

Each position belongs to a separate Arena category. These are human preference rankings—not hardware recommendations. We do not combine the categories into one score.

Agents#5of 23 open models

Kimi K3 (Max)moonshotLicense: Kimi K3 license

Export dated Nov 4, 2026Checked: Nov 7, 2026Overall board: #25

View official Agents board

Web coding#3of 62 open models

Kimi K3 (Max)moonshotLicense: Kimi K3 license

Export dated Nov 6, 2026Checked: Nov 7, 2026Overall board: #18

View official Web coding board

Text#1of 228 open models

Kimi K3 (Max)moonshotLicense: Kimi K3 license

Export dated Nov 6, 2026Checked: Nov 7, 2026Overall board: #19

View official Text board

Rankings reflect the dated export; live Arena boards can change. Open-source positions use Arena’s open-model filter. The overall position also includes the other models on that category board. Availability of weights and license terms still need to be checked separately.

What do these scores mean?

Two ways to compare models. Neither score measures answer quality, benchmark performance, or whether a model fits your machine.

Freshness · release age

A new release starts at 100. The score halves every 180 days: 100 × 2^(−age in days / 180). When the release date is unavailable, a repository date is a fallback—not proof of when the model was released. Missing dates stay unscored.

Check the date source

Popularity · adoption

Downloads over the past 30 days carry 80% of the score; lifetime likes carry 20%. Both use a logarithmic scale, capped at 10 million downloads and 100,000 likes. If likes are unavailable, downloads carry all the weight. Missing download statistics stay unscored.

round(80 × min(1, log10(1 + downloads) / 7) + 20 × min(1, log10(1 + likes) / 5))

Check the statistics source

Scores as of Nov 8, 2026. Statistics checked Nov 7, 2026. Statistics are marked for refresh after 72 hours.Full scoring method

At a glance

Parameters: 2.8T total · 104B active
Architecture: Mixture of experts
Context length: 1,000,000
License: Kimi K3 License
Software: vLLM, SGLang, Transformers

View the model source

MY HARDWARE

Will it run on your machine?

Save your machine to see a personalized rating and its reasoning.

Add my hardware

Tradeoffs

Publisher repository exceeds 1 TB; confirm cluster memory, download time, and cost before renting.

Ways to run it

vLLM · Linux

The publisher repository exposes the serving example, but the weight files are over 1 TB. This requires a large multi-GPU cluster rather than an ordinary single rented GPU.

vllm serve "moonshotai/Kimi-K3"

Official instructions

Explore its uses

AGENTSCODINGMULTIMODALREASONING

Keep exploring

Agents

Z.ai

GLM 5.3

Explore 2 versions of GLM 5.3. Compare model sizes, packages, hardware requirements, and available setups.

2 versionsCompare sizes & formats

Gemma 4

Explore 4 versions of Gemma 4. Compare model sizes, packages, hardware requirements, and available setups.

4 versionsCompare sizes & formats

Qwen3.5

Explore 3 versions of Qwen3.5. Compare model sizes, packages, hardware requirements, and available setups.

3 versionsCompare sizes & formats

LOCALRENTED GPU

Compare versions →