[whatmodel.sh](../../README.md) / Mistral Small Instruct 22B (2409) # Mistral Small Instruct 22B (2409) Mistral AI · released 2024-09 · 22.25B params · mistral-research **score 50.5** ±1.7 · #358 of 732 · from 9 benchmark results across 6 sources Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only. ## What it's good at _{● marks Mistral Small Instruct 22B (2409) among every model on that task or test, worst to best.}

task         vs the others          rank  counts  tests
agent tasks  ▁▁▃●█▄▅▅▄▄▂▁   #269 of 313*    5.8%      1
coding       ▁▁▁▃▅▇●▇▅▄▃▁    #245 of 469     28%      4

agent tasks · #269 of 313 · 5.8% of its score

Beats [DeepSeek R1 0528 Qwen3 8B](deepseek-r1-0528-qwen3-8b.md), [Magistral Small 1.0 (2506)](magistral-small-2506.md), [Llama 4 Scout (109B-A17B)](llama-4-scout.md) here, though they rank higher overall. Trails [Granite 4.0 H 1B](granite-4-0-h-1b.md), [Granite 4.0 1B](granite-4-0-1b.md), [SmolLM3 3B](smollm3-3b.md) here, though they rank lower overall.

test                                result  vs the others in its run  difficulty      counts  why
Galileo Agent Leaderboard v1 (too…   75.0%  ▃▃●▃▆▅▆▅█▂▅▂ beats 5/31   easy · avg 78%    5.8%  typical

coding · #245 of 469 · 28% of its score

Beats [Qwen3-VL 32B Instruct](qwen3-vl-32b-instruct.md), [Solar Open 100B (102B-A12B)](solar-open-100b.md), [GLM-4.7-Flash](glm-4-7-flash.md) here, though they rank higher overall. Trails [QwQ-32B-Preview](qwq-32b-preview.md), [InternLM2.5 20B Chat](internlm2-5-20b-chat.md), [Ministral 3 3B Reasoning 2512](ministral-3-3b-reasoning.md) here, though they rank lower overall.

test                                result  vs the others in its run  difficulty        counts  why
CodeReviewQA: change localisation…   31.6%  █▂▁▁▂▂▂▃●▅▅▇ beats 27/47  hard · avg 33%       11%  sharp · at its level · noisy
CodeReviewQA: solution identifica…   60.1%  ▃··▁▂▂▃▂█●▄▂ beats 34/47  medium · avg 51%    9.3%  sharp · at its level · noisy
CodeReviewQA: automated code refi…   45.6%  ▁▁▁▁▁▁▂▂▃█▅● beats 43/47  hard · avg 32%      7.7%  at its level
CodeReviewQA: change type recogni…   76.7%  ▁▁▁······▁▃● beats 36/47  easy · avg 67%      0.3%  blunt · at its level · noisy

other tests (4)

test                                result  vs the others in its run  difficulty        counts  why
Polish EQ-Bench v2 (SpeakLeash)      72.9%  ▁··▁▁▂▂▆▅▆█● beats 62/70  medium · avg 62%     17%  at its level
EvalPerf correctness (118 tasks, …   69.5%  ▂·▂▂▂●▅▆▆▆▃█ beats 6/29   easy · avg 74%       17%  typical
BRIDGE clinical text, zero-shot a…   35.2%  ▁▁▁▁▁▂▂▃▄●█▃ beats 46/97  hard · avg 31%       17%  blunt · at its level
Open Persian LLM Leaderboard v2 a…   44.0%  ▂▄▂●█▄▃▆▃▅▂▂ beats 17/81  medium · avg 57%     14%  blunt · 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
CodeReviewQA: The Code Review Comprehen…        4   28% ▇▇▇
SpeakLeash Polish EQ-Bench leaderboard          1   17% ▇▇
EvalPlus / EvalPerf published evaluatio…        1   17% ▇▇
BRIDGE Medical Leaderboard                      1   17% ▇▇
Open Persian LLM Leaderboard v2                 1   14% ▇▇
Galileo Agent Leaderboard                       1    6% ▇
prior: pull toward the middle                        1% ▇
weighted average                                   = 50.5

_{Licences: CodeReviewQA: The Code Review Comprehension Assessment for Large Language Models ([CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/)) · SpeakLeash Polish EQ-Bench leaderboard ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · EvalPlus / EvalPerf published evaluation runs ([Apache-2.0](../../licenses/evalplus-published-runs-Apache-2.0.txt)) · BRIDGE Medical Leaderboard ([MIT](../../licenses/bridge-medical-leaderboard-MIT.txt)) · Open Persian LLM Leaderboard v2 ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · Galileo Agent Leaderboard ([Apache-2.0](../../licenses/galileo-agent-leaderboard-Apache-2.0.txt))} ## How to run it ### 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 Mistral Small Instruct 22B (2409) (Q4_K_M, needs ~13.3 GB, from [huggingface.co/lmstudio-community](https://huggingface.co/lmstudio-community/Mistral-Small-Instruct-2409-GGUF/tree/9e62590543cada96cace75d7b79616ce7f194c41)) ```sh llama cli -hf lmstudio-community/Mistral-Small-Instruct-2409-GGUF:Q4_K_M ``` 3. Or start a local server and open the address it prints (http://localhost:9931) ```sh llama serve -hf lmstudio-community/Mistral-Small-Instruct-2409-GGUF:Q4_K_M ``` Other versions: Q8_0 (~23.6 GB) `llama cli -hf lmstudio-community/Mistral-Small-Instruct-2409-GGUF:Q8_0` ### 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 Mistral Small Instruct 22B (2409) (Q4_K_M, needs ~13.3 GB, from [huggingface.co/lmstudio-community](https://huggingface.co/lmstudio-community/Mistral-Small-Instruct-2409-GGUF/tree/9e62590543cada96cace75d7b79616ce7f194c41)) ```sh lms get https://huggingface.co/lmstudio-community/Mistral-Small-Instruct-2409-GGUF@Q4_K_M -y ``` 3. Then run lms ls to see its name, and lms chat followed by that name Other versions: Q8_0 (~23.6 GB) `lms get https://huggingface.co/lmstudio-community/Mistral-Small-Instruct-2409-GGUF@Q8_0 -y` · 4bit (~12.5 GB, Mac only) `lms get https://huggingface.co/mlx-community/Mistral-Small-Instruct-2409-4bit -y` · 8bit (~23.6 GB, Mac only) `lms get https://huggingface.co/mlx-community/Mistral-Small-Instruct-2409-8bit -y` ### Ollama 1. Install Ollama. [Download it (macOS / Windows)](https://ollama.com/download) ```sh # Linux curl -fsSL https://ollama.com/install.sh | sh ``` 2. Run Mistral Small Instruct 22B (2409) (Q4_K_M, needs ~13.85 GB, from [ollama.com](https://ollama.com/library/mistral-small/tags)) ```sh ollama run mistral-small:22b-instruct-2409-q4_K_M ``` Other versions: Q8_0 (~27.2 GB) `ollama run mistral-small:22b-instruct-2409-q8_0` ### 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 Mistral Small Instruct 22B (2409) (4bit, needs ~12.5 GB, Mac only, from [huggingface.co/mlx-community](https://huggingface.co/mlx-community/Mistral-Small-Instruct-2409-4bit/tree/81b15f31a74f2ed106aa142e61454e3d478db0ce)) ```sh mlx_lm.chat --model mlx-community/Mistral-Small-Instruct-2409-4bit ``` 3. Or start an OpenAI-compatible server at http://localhost:8080/v1 ```sh mlx_lm.server --model mlx-community/Mistral-Small-Instruct-2409-4bit ``` Other versions: 8bit (~23.6 GB, Mac only) `mlx_lm.chat --model mlx-community/Mistral-Small-Instruct-2409-8bit` Model repository and documentation: [huggingface.co/mistralai](https://huggingface.co/mistralai/Mistral-Small-Instruct-2409) [interactive version](https://whatmodel.sh/model/mistral-small-2409) · [JSON](https://whatmodel.sh/api?format=json&all=1)