[whatmodel.sh](../../README.md) / Kimi K2.7 Code # Kimi K2.7 Code Moonshot AI · released 2026-06 · 1000B params (32B active) · modified-mit **score 70.8** ±0.9 · #73 of 732 · from 7 benchmark results across 2 sources Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only. ## What it's good at ● marks Kimi K2.7 Code among every model on that task or test, worst to best. 
task    vs the others         rank  counts  tests
coding  ▁▁▁▃▅▇█▇▅●▃▁   #62 of 469*     15%      1
math    ▁▁▃▅█▆▄█●▂▁▁    #56 of 431     67%      3
coding · #62 of 469 · 15% of its score Beats [GPT-5.4](gpt-5-4.md), [Gemini 3.1 Pro (Preview)](gemini-3-1-pro.md), [DeepSeek V4 Flash Vision (Exp)](deepseek-v4-flash-vision-exp.md) here, though they rank higher overall. Trails [MiMo-V2.5-Pro (1T-A42B)](mimo-v2-5-pro.md), [MiniMax-M3](minimax-m3.md), [GLM-5.1](glm-5-1.md) here, though they rank lower overall. 
test                       result  vs the others in its run   difficulty  counts  why
Arena WebDev (Code Arena)  1473.5  ▁▁▃▂▅█●▆▅▅▁▂ beats 67/129  avg 981        15%  sharp
math · #56 of 431 · 67% of its score Beats [Qwen3.6 Max Preview](qwen3-6-max-preview.md), [Claude Sonnet 4.6](claude-sonnet-4-6.md), [DeepSeek V4 Pro (Preview, Apr 2026)](deepseek-v4-pro-preview.md) here, though they rank higher overall. Trails [A.X K2 (688B-A33B)](ax-k2.md), [DeepSeek V4 Flash (Preview, Apr 2026)](deepseek-v4-flash-preview.md) here, though they rank lower overall. 
test                       result  vs the others in its run    difficulty            counts  why
FrontierMath tiers 1–3 v2   54.0%  ▂··▁▁▃▄▄●▆▄▃ beats 38/86    very hard · avg 1.2%     42%  sharp · at its level
FrontierMath tier 4 v2      12.2%  ▁▂▁●█▅▂▃▂▁▁▁ beats 12/67    very hard · avg 0.1%     23%  sharp
OTIS mock AIME 2024–2025    95.6%  ▃▃▄▃▃▃▇▇█●▂▅ beats 162/197  very hard · avg 7.2%    2.4%  sharp · noisy
other tests (3) 
test                   result  vs the others in its run    difficulty            counts  why
GPQA Diamond            87.9%  ▁▁▁▁▂▄▆▄▅▇●▅ beats 160/211  hard · avg 39%          8.4%  sharp
Chess Puzzles (Epoch)   21.0%  █▁▁▁▁▂●▃▃▂▁▁ beats 84/137   very hard · avg 5.8%    6.1%  sharp · noisy
SimpleQA Verified       36.5%  ▁▁▃▂▃▃●▆█▄▂▅ beats 31/81    very hard · avg 8.0%    3.2%  at its level
## 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
Epoch AI                             6   84% ▇▇▇▇▇▇▇▇▇▇
LMArena leaderboard dataset          1   15% ▇▇
prior: pull toward the middle             0% ▇
weighted average                        = 70.8
Licences: Epoch AI ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · LMArena leaderboard dataset ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) ## How to run it ### Through an API api (Moonshot AI) · $1.71/M blended · $0.95 in / $4 out per 1M tokens · [platform.kimi.ai](https://platform.kimi.ai/docs/guide/kimi-k2-7-code-quickstart) [price source](https://platform.kimi.ai/docs/pricing/chat.md): USD per 1M tokens, Kimi API (Moonshot AI) pay-as-you-go, kimi-k2.7-code: input (cache miss) $0.95, output $4.00; cache-hit input $0.19 and kimi-k2.7-code-highspeed ($1.90/$8.00) excluded; taxes excluded. Currency: USD. Moonshot AI Kimi official pricing page, observed 2026-09-25. · [context source](https://platform.kimi.ai/docs/pricing/chat.md): Moonshot AI Kimi API pricing page: kimi-k2.7-code context window 262,144 tokens (the model card at @74797c9 states 'Context Length 256K'). API serving limit of the maker's own service. Observed 2026-09-25. ### 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 K2.7 Code (ud-q4_k_xl, needs ~584.66 GB, from [huggingface.co/unsloth](https://huggingface.co/unsloth/Kimi-K2.7-Code-GGUF/tree/46352ca8dc32aa60f9754f5bd3fb778deeb9e430)) ```sh llama cli -hf unsloth/Kimi-K2.7-Code-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-K2.7-Code-GGUF:UD-Q4_K_XL ``` Other versions: ud-q2_k_xl (~340.42 GB) `llama cli -hf unsloth/Kimi-K2.7-Code-GGUF:UD-Q2_K_XL` ### 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 K2.7 Code (ud-q4_k_xl, needs ~584.66 GB, from [huggingface.co/unsloth](https://huggingface.co/unsloth/Kimi-K2.7-Code-GGUF/tree/46352ca8dc32aa60f9754f5bd3fb778deeb9e430)) ```sh lms get https://huggingface.co/unsloth/Kimi-K2.7-Code-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 (~340.42 GB) `lms get https://huggingface.co/unsloth/Kimi-K2.7-Code-GGUF@UD-Q2_K_XL -y` Model repository and documentation: [huggingface.co/moonshotai](https://huggingface.co/moonshotai/Kimi-K2.7-Code) [interactive version](https://whatmodel.sh/model/kimi-k2-7-code) · [JSON](https://whatmodel.sh/api?format=json&all=1)