[whatmodel.sh](../../README.md) / Kimi K2 Instruct (0711) # Kimi K2 Instruct (0711) Moonshot AI · released 2025-07 · 1000B params (32B active) · modified-mit **score 65.8** ±0.8 · #129 of 732 · from 29 benchmark results across 9 sources Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only. ## What it's good at _{● marks Kimi K2 Instruct (0711) among every model on that task or test, worst to best.}

task            vs the others          rank  counts  tests
agent tasks     ▁▁▃▆█▄●▅▄▄▂▁    #125 of 313     26%      7
coding          ▁▁▁▃▅▇█●▅▄▃▁    #138 of 469     31%      5
chat & writing  ▁▁▁▃▆▄▆█▆●▇▃    #94 of 402*    8.8%      2
math            ▁▁▃▅█▆▄●▄▂▁▁   #134 of 431*    4.0%      2

agent tasks · #125 of 313 · 26% of its score

Beats [o3-pro](o3-pro.md), [o3](o3.md), [GPT-5 mini](gpt-5-mini.md) here, though they rank higher overall. Trails [SmolLM3 3B](smollm3-3b.md), [Jamba 1.5 Large](jamba-1-5-large.md), [Devstral 2 (2512)](devstral-2.md) here, though they rank lower overall.

test                                result  vs the others in its run   difficulty            counts  why
τ²-bench Telecom                     67.5%  ██··███···●█ beats 5/6     very hard · avg 8.9%    9.5%  sharp · at its level · several runs
SWE-bench Verified                   64.6%  ▁▁▁▁▁▁▂▄●▆▁▁ beats 66/138  hard · avg 17%          5.4%  sharp · at its level
τ²-bench Retail                      70.8%  ▅···●·▃····▃ beats 3/6     medium · avg 40%        3.5%  at its level
BFCL V4 (December 2025 archive)      59.1%  ▁▁▁▂██▅▅▆●▂▂ beats 70/77   hard · avg 31%          3.0%  at its level
τ²-bench Airline                     54.2%  ▄·▄··▄···●▄▄ beats 3/6     hard · avg 25%          2.3%  at its level
Terminal-Bench (pre-2.0 release)     25.9%  ▄·▄···●▄···█ beats 2/5     very hard · avg 12%     1.4%  at its level
Galileo Agent Leaderboard v2 (act…   53.0%  ▅▃·▃▃▅▅█▃●▅▅ beats 15/19   hard · avg 26%          1.3%  at its level

coding · #138 of 469 · 31% of its score

Beats [Qwen3.5 Plus](qwen3-5-plus.md), [Qwen3-Coder-Next (80B-A3B)](qwen3-coder-next.md), [Gemini 2.5 Pro](gemini-2-5-pro.md) here, though they rank higher overall. Trails [Llama 3.1 Nemotron Ultra 253B v1](llama-3-1-nemotron-ultra-253b.md), [Gemma 4 12B](gemma-4-12b.md), [Seed-OSS 36B Instruct](seed-oss-36b.md) here, though they rank lower overall.

test                                result  vs the others in its run   difficulty            counts  why
Aider polyglot (225 exercises, tw…   59.1%  ▂·▂▃▅▂▃█●▄▂▂ beats 33/49   very hard · avg 7.6%     15%  sharp · at its level
SWE-bench Verified                   64.6%  ▁▁▁▁▁▁▂▄●▆▁▁ beats 66/138  hard · avg 17%          5.4%  sharp · at its level
LiveCodeBench (version unspecifie…   46.7%  █·······▄▄●▄ beats 4/5     hard · avg 21%          4.4%  at its level
MBPP+                                81.8%  ▃·····▃█▃··● beats 6/6     medium · avg 60%        3.1%  typical
HumanEval+                           86.0%  ●▄▄·····▄▄·█ beats 0/6     easy · avg 66%          2.8%  sharp · maxed out

chat & writing · #94 of 402 · 8.8% of its score

Beats [LongCat 2.0 (1.6T-A48B)](longcat-2-0.md), [Inkling Small (266B-A12B)](inkling-small.md), [Qwen3.5 122B-A10B](qwen3-5-122b-a10b.md) here, though they rank higher overall. Trails [Gemma 4 12B](gemma-4-12b.md), [Kimi K2 Instruct (0905)](kimi-k2-0905.md), [Gemma 4 26B-A4B](gemma-4-26b-a4b.md) here, though they rank lower overall.

test                              result  vs the others in its run    difficulty      counts  why
LMArena (text)                    1418.2  ▁▁▁▂▃▃▃▆▆●█▃ beats 270/385  avg 1260          6.4%  blunt
IFEval (aggregation unspecified)   88.9%  █··█·█··██●█ beats 5/6      easy · avg 74%    2.4%  typical

math · #134 of 431 · 4.0% of its score

Beats [Claude Sonnet 4.6](claude-sonnet-4-6.md), [MiniMax-M3](minimax-m3.md), [MiMo-V2.5 (310B-A15B)](mimo-v2-5.md) here, though they rank higher overall. Trails [Olmo 3.1 32B Think](olmo-3-1-32b-think.md), [Qwen3-VL 30B-A3B Thinking](qwen3-vl-30b-a3b-thinking.md), [Granite 4.2 8B](granite-4-2-8b.md) here, though they rank lower overall.

test       result  vs the others in its run  difficulty           counts  why
AIME 2025   50.7%  ▄▄···▄●··▄·█ beats 3/6    hard · avg 16%         3.0%  sharp · at its level · noisy
MATH-500    97.6%  ▄·▄··█··●·▄▄ beats 4/6    very easy · avg 87%    1.1%  maxed out · noisy

other tests (13)

test                                result  vs the others in its run    difficulty            counts  why
MMLU                                 89.9%  ▄·····█●▄▄·▄ beats 3/6      medium · avg 68%        6.6%  typical
GPQA Diamond                         75.8%  █·█·█··●█··█ beats 3/5      hard · avg 40%          6.4%  sharp · at its level
MMLU-Pro                             82.1%  █●·▄···▄··▄▄ beats 2/6      medium · avg 58%        5.0%  typical
EQ-Bench 3 (emotional intelligenc…  1311.0  ▁·▂▁▄▅▅▃█●▁▁ beats 52/62    avg 1034                3.9%  typical
CMMLU                                89.7%  █····▄··▄▄▄● beats 6/6      medium · avg 67%        2.7%  typical
C-Eval                               91.3%  █····▄·▄·▄●▄ beats 5/6      medium · avg 68%        2.4%  typical
VitaBench                            18.2%  █·······●█·█ beats 2/6      very hard · avg 6.5%    2.4%  typical
COLLIE (constrained text generati…   56.3%  ▄···▄█▄····● beats 5/6      hard · avg 30%          2.4%  at its level
ZebraLogic                           89.1%  █···█··█●█·█ beats 3/5      medium · avg 51%        1.0%  noisy
BeyondAIME                           36.6%  █····●·▄▄··▄ beats 3/6      hard · avg 24%          0.8%  at its level
Open PT-LLM Leaderboard average (…   81.8%  ▃▁▁▁▁▁▁▁▂█▅● beats 266/276  medium · avg 72%        0.8%  noisy
DROP                                 89.0%  ▄···▄·▄█·▄·● beats 6/6      medium · avg 58%        0.3%  blunt · at its level · noisy
Vectara hallucination leaderboard…   98.9%  ▁▁▂▃▄▄█▆▃▂●▁ beats 137/140  very easy · avg 94%     0.2%  blunt · maxed out · noisy

## How the score was worked out Each result implies a score on its own, after adjusting for how hard that test is and who ran it. The model's 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 is added up by who ran the tests. [Method](../method.md)

source                                    results  share of the score
Meituan LongCat-Flash-Chat model card          21   62% ▇▇▇▇▇▇▇
Aider polyglot benchmark runs                   1   15% ▇▇
LMArena leaderboard dataset                     1    6% ▇
Moonshot AI Kimi-K2-Instruct model card…        1    6% ▇
EQ-Bench 3 canonical leaderboard results        1    4% ▇
Berkeley Function Calling Leaderboard r…        1    3% ▇
Galileo Agent Leaderboard                       1    1% ▇
Open Portuguese LLM Leaderboard                 1    1% ▇
Vectara hallucination leaderboard               1    0% ▇
prior: pull toward the middle                        0% ▇
weighted average                                   = 65.8

_{Licences: Meituan LongCat-Flash-Chat model card ([MIT](https://huggingface.co/meituan-longcat/LongCat-Flash-Chat/blob/f0b8e8d7380330f450ca1fde4ec56823a514584c/LICENSE)) · Aider polyglot benchmark runs ([Apache-2.0](../../licenses/aider-polyglot-runs-Apache-2.0.txt)) · LMArena leaderboard dataset ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · Moonshot AI Kimi-K2-Instruct model card ([Modified MIT License (Kimi K2)](https://huggingface.co/moonshotai/Kimi-K2-Instruct/blob/fd1984e2b7a3350dbf7305fe73a4ede25c14de50/LICENSE)) · EQ-Bench 3 canonical leaderboard results ([MIT](../../licenses/eqbench3-MIT.txt)) · Berkeley Function Calling Leaderboard result archive ([Apache-2.0](../../licenses/bfcl-result-archive-v4-Apache-2.0.txt)) · Galileo Agent Leaderboard ([Apache-2.0](../../licenses/galileo-agent-leaderboard-Apache-2.0.txt)) · Open Portuguese LLM Leaderboard ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · Vectara hallucination leaderboard ([Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0))}

Not counted (3)

source                  test              score  why
τ²-bench leaderboard:…  τ²-bench Retail   70.6%  only model in this source's run
τ²-bench leaderboard:…  τ²-bench Airline  56.5%  only model in this source's run
τ²-bench leaderboard:…  τ²-bench Telecom  65.8%  only model in this source's run

## How to run it ### Through an API api (Novita AI) · $1.00/M blended · $0.57 in / $2.3 out per 1M tokens · [novita.ai](https://novita.ai/models) _{[price source](https://github.com/BerriAI/litellm/blob/c19ce71bcdad234ca87a7af10cee45375ce811a8/model_prices_and_context_window.json): USD per 1M text tokens, standard uncached input and output. Model endpoint moonshotai/kimi-k2-instruct. July 2025 Kimi K2 Instruct endpoint; not K2-0905 or K2 Thinking. Product: Serverless chat completions. API model: moonshotai/kimi-k2-instruct. Settings: service tier=standard, batch=no, cache=uncached input. Additional/excluded charges: Tools, cache storage, priority service and other modalities excluded.. Currency: USD. · [context source](https://github.com/BerriAI/litellm/blob/c19ce71bcdad234ca87a7af10cee45375ce811a8/model_prices_and_context_window.json): provider maximum input tokens; not an intrinsic model maximum or legacy max\_tokens} ### llama.cpp (recommended) 1. Install llama.cpp. Homebrew and winget installs name the commands llama-cli and llama-server instead. [Other ways to install](https://github.com/ggml-org/llama.cpp/blob/master/docs/install.md) ```sh # macOS / Linux curl -LsSf https://llama.app/install.sh | sh # Windows irm https://llama.app/install.ps1 | iex ``` 2. Run Kimi K2 Instruct (0711) (ud-q4_k_xl, needs ~600.5 GB, from [huggingface.co/unsloth](https://huggingface.co/unsloth/Kimi-K2-Instruct-GGUF/tree/main)) ```sh llama cli -hf unsloth/Kimi-K2-Instruct-GGUF:UD-Q4_K_XL ``` 3. Or start a local server and open the address it prints (http://localhost:9931) ```sh llama serve -hf unsloth/Kimi-K2-Instruct-GGUF:UD-Q4_K_XL ``` Other versions: ud-q2_k_xl (~300.5 GB) `llama cli -hf unsloth/Kimi-K2-Instruct-GGUF:UD-Q2_K_XL` · ud-iq1_s (~150.5 GB) `llama cli -hf unsloth/Kimi-K2-Instruct-GGUF:UD-IQ1_S` ### LM Studio 1. Install LM Studio. Open the app once before using the lms command. [Download it (macOS / Windows / Linux)](https://lmstudio.ai/download) 2. Run Kimi K2 Instruct (0711) (ud-q4_k_xl, needs ~600.5 GB, from [huggingface.co/unsloth](https://huggingface.co/unsloth/Kimi-K2-Instruct-GGUF/tree/main)) ```sh lms get https://huggingface.co/unsloth/Kimi-K2-Instruct-GGUF@UD-Q4_K_XL -y ``` 3. Then run lms ls to see its name, and lms chat followed by that name Other versions: ud-q2_k_xl (~300.5 GB) `lms get https://huggingface.co/unsloth/Kimi-K2-Instruct-GGUF@UD-Q2_K_XL -y` · ud-iq1_s (~150.5 GB) `lms get https://huggingface.co/unsloth/Kimi-K2-Instruct-GGUF@UD-IQ1_S -y` · 4bit (~577.6 GB, Mac only) `lms get https://huggingface.co/mlx-community/Kimi-K2-Instruct-4bit -y` ### MLX (Mac) 1. Install MLX (Mac). Apple silicon Macs only, text only. ```sh # with uv uv tool install mlx-lm # or pip install mlx-lm ``` 2. Run Kimi K2 Instruct (0711) (4bit, needs ~577.6 GB, Mac only, from [huggingface.co/mlx-community](https://huggingface.co/mlx-community/Kimi-K2-Instruct-4bit/tree/91fb4f9fd1de100104925196d62b8ee06fd2ad60)) ```sh mlx_lm.chat --model mlx-community/Kimi-K2-Instruct-4bit ``` 3. Or start an OpenAI-compatible server at http://localhost:8080/v1 ```sh mlx_lm.server --model mlx-community/Kimi-K2-Instruct-4bit ``` Model repository and documentation: [huggingface.co/moonshotai](https://huggingface.co/moonshotai/Kimi-K2-Instruct) [interactive version](https://whatmodel.sh/model/kimi-k2) · [JSON](https://whatmodel.sh/api?format=json&all=1)