whatmodel.sh / Kimi K2 Instruct (0905)

Kimi K2 Instruct (0905)

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Kimi K2 Instruct (0905) is an open-weights language model from Moonshot AI, released September 2025. Its published size is 1000B parameters (32B active per token). Licence: modified-mit.

Catalogue identity notes

Comparison-only. Arena entry 'kimi-k2-0905-preview' (text). Maker repo moonshotai/Kimi-K2-Instruct-0905 @ac6c49f: licence tag other (modified-mit). Parameters: card: 'Total Parameters 1T', 'Activated Parameters 32B'. Released 2025-09: Kimi platform changelog, September 2025: 'Launched the kimi-k2-0905-preview model'. Arena 'kimi-k2-0905-preview' is the Kimi API name for Kimi-K2-Instruct-0905 (same '-preview' API naming as the catalogued kimi-k2 / kimi-k2-0711-preview). Kimi changelog, May 2026: the kimi-k2 series including kimi-k2-0905-preview 'was retired'.

63.8±2.4#159 of 719 LLMsfrom 5 benchmark results across 4 sources

$ how it compares

The LLMs ranked either side of Kimi K2 Instruct (0905), out of 828 with a score, then more from Moonshot AI.

#modelscorewhereRAMreleased
180Mistral Medium 3.1 (2508) Mistral64±2comparison only2025-08
181Gemma 4 26B-A4B Google64±1api $0.14/M +local~18GB2026-04
182Kimi K2 Instruct (0905) Moonshot AI64±2local~298GB2025-09
183Qwen3-VL 235B-A22B Instruct Alibaba64±1local~134GB2025-09
184Nova 2 Pro (Preview) Amazon64±2api $3.78/M2025-11
more from Moonshot AI
25Kimi K3 Moonshot AI77±1api $6.00/M +local~884GB2026-07
61Kimi K2.6 Moonshot AI73±1api $1.71/M +local~317GB2026-04
72Kimi K2.7 Code Moonshot AI72±1api $1.71/M +local~359GB2026-06

see all LLMs · find the ones that fit your machine

$ what it's good at

coding▁▁▁▃▆██▇▆▄▃▁ #160 of 458 * · 24% of its score · 1 test

Beats Gemini 2.5 Pro, Qwen3-Coder-Next (80B-A3B), Grok 4.1 Fast (Reasoning) here, though they rank higher overall. Trails QwQ-32B-Preview, Ministral 3 8B Reasoning 2512, Phi-4-reasoning 14B here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
WebApp1K-Duo-React pass@1 (onekq-ai)57.9% ✎▁··▁·▁▁▁▆█▇█ beats 11/45hard · avg 32%▇▇▇▇▇▇▇▇▇▇ 24%at its level
chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #89 of 393 * · 50% of its score · 1 test

Beats LongCat 2.0 (1.6T-A48B), Inkling Small (266B-A12B), Qwen3.5 122B-A10B here, though they rank higher overall. Trails Gemma 4 12B here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
LMArena (text)1418.7 ✎▁▁▁▂▃▃▃▆▆▇█▃ beats 269/379avg 1268▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 50%blunt
other tests3 not in a task · 23% of its score
testresultvs the others in its rundifficultycountswhy
CPTUB average (SpeakLeash, 0-5 judge scale)79.7% ✎▁▂▂▂▃█▃▆▅▄▂▃ beats 80/95easy · avg 66%▇▇▇▇▇▇▇▇▇ 22%typical
Vectara hallucination leaderboard, 2025 dataset (HHEM-2.3, factual consistency rate)93.8% ✎▁▁▂▃▄▄█▆▃▂▂▁ beats 50/140very easy · avg 94%▇ 1.6%blunt · maxed out · noisy
Vectara hallucination leaderboard (HHEM-2.3, factual consistency rate)82.1% ✎▁▃▁▂▅█▂▃▄▃▁▁ beats 8/98very easy · avg 90%▇ 0.0%blunt

$ where to run it

llama.cpp · recommended

version

  1. Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.
    macOS / Linux
    curl -LsSf https://llama.app/install.sh | sh
    Windows
    irm https://llama.app/install.ps1 | iex

    other ways to install ↗

  2. Run Kimi K2 Instruct (0905) · needs ~588.12 GB · from huggingface.co/unsloth
    chat in the terminal
    llama cli -hf unsloth/Kimi-K2-Instruct-0905-GGUF:UD-Q4_K_XL
  3. Or start a local server and open the address it prints (http://localhost:9931)
    llama serve -hf unsloth/Kimi-K2-Instruct-0905-GGUF:UD-Q4_K_XL

LM Studio

version

  1. Install LM Studio · Open the app once before using the lms command.

    download it (macOS / Windows / Linux) ↗

  2. Run Kimi K2 Instruct (0905) · needs ~588.12 GB · from huggingface.co/unsloth

    open in LM Studio ↗

  3. Or download it in the terminal
    lms get https://huggingface.co/unsloth/Kimi-K2-Instruct-0905-GGUF@UD-Q4_K_XL -y
  4. Then run lms ls to see its name, and lms chat followed by that name

MLX (Mac)

  1. Install MLX (Mac) · Apple silicon Macs only, text only.
    with uv
    uv tool install mlx-lm
    or
    pip install mlx-lm
  2. Run Kimi K2 Instruct (0905) · needs ~449.3 GB, Mac only · from huggingface.co/mlx-community
    chat in the terminal
    mlx_lm.chat --model mlx-community/Kimi-K2-Instruct-0905-mlx-3bit
  3. Or start an OpenAI-compatible server at http://localhost:8080/v1
    mlx_lm.server --model mlx-community/Kimi-K2-Instruct-0905-mlx-3bit

Official weights and documentation: huggingface.co/moonshotai ↗

$ how the score was worked out · LMArena leaderboard dataset 50%, WebApp1K models leaderboard (… 24%, SpeakLeash CPTUB (Complex Pol… 22% and 1 more

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
LMArena leaderboard dataset1▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 50%CC BY 4.0
WebApp1K models leaderboard (onekq-ai)1▇▇▇▇▇▇▇ 24%MIT
SpeakLeash CPTUB (Complex Polish Text Understanding Benchmark)1▇▇▇▇▇▇ 22%MIT
Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository)2▇ 2%Apache-2.0
prior · pull toward the middle▇ 2%
weighted average= 63.8

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json