Mistral Small 4 is a powerful hybrid model capable of acting as both a general instruction model and a reasoning model. It unifies the capabilities of three different model families—Instruct, Reasoning (previously called Magistral), and Devstral—into a single, unified model.
With its multimodal capabilities, efficient architecture, and flexible mode switching, it is a powerful general-purpose model for any task. In a latency-optimized setup, Mistral Small 4 achieves a 40% reduction in end-to-end completion time, and in a throughput-optimized setup, it handles 3x more requests per second compared to Mistral Small 3.
To further improve efficiency you can either take advantages of:
Mistral Small 4 is designed for general chat assistants, coding, agentic tasks, and reasoning tasks (with reasoning mode toggled). Its multimodal capabilities also enable document and image understanding for data extraction and analysis.
Its capabilities are ideal for:
Developers interested in coding and agentic capabilities for SWE automation and codebase exploration.
Enterprises seeking general chat assistants, agents, and document understanding.
Researchers leveraging its math and research capabilities.
Mistral Small 4 is also well-suited for customization and fine-tuning for more specialized tasks.
reasoning_effort="high": Deep, step-by-step reasoning for complex problems, with equivalent verbosity to previous Magistral models such as mistralai/Magistral-Small-2509.
Mistral Small 4 with reasoning achieves competitive scores, matching or surpassing GPT-OSS 120B across all three benchmarks while generating significantly shorter outputs. On AA LCR, Mistral Small 4 scores 0.72 with just 1.6K characters, whereas Qwen models require 3.5-4x more output (5.8-6.1K) for comparable performance. On LiveCodeBench, Mistral Small 4 outperforms GPT-OSS 120B while producing 20% less output. This efficiency reduces latency, inference costs, and improves user experience.
You can find Mistral Small 4 support on multiple libraries for inference and fine-tuning. We here thank everyone contributors and maintainers that helped us making it happen.
Tip. Use our custom Docker image with fixes for tool calling and reasoning parsing in vLLM, and the latest Transformers version. We are working with the vLLM team to merge these fixes soon.
Mistral Small 4 can follow your instructions to the letter.
from datetime import datetime, timedelta
from openai import OpenAI
from huggingface_hub import hf_hub_download
# Modify OpenAI's API key and API base to use vLLM's API server.
openai_api_key = "EMPTY"
openai_api_base = "http://localhost:8000/v1"
TEMP = 0.1
client = OpenAI(
api_key=openai_api_key,
base_url=openai_api_base,
)
models = client.models.list()
model = models.data[0].id
def load_system_prompt(repo_id: str, filename: str) -> str:
file_path = hf_hub_download(repo_id=repo_id, filename=filename)
with open(file_path, "r") as file:
system_prompt = file.read()
today = datetime.today().strftime("%Y-%m-%d")
yesterday = (datetime.today() - timedelta(days=1)).strftime("%Y-%m-%d")
model_name = repo_id.split("/")[-1]
return system_prompt.format(name=model_name, today=today, yesterday=yesterday)
SYSTEM_PROMPT = load_system_prompt(model, "SYSTEM_PROMPT.txt")
messages = [
{"role": "system", "content": SYSTEM_PROMPT},
{
"role": "user",
"content": "Write me a sentence where every word starts with the next letter in the alphabet - start with 'a' and end with 'z'.",
},
]
response = client.chat.completions.create(
model=model,
messages=messages,
temperature=TEMP,
reasoning_effort="none",
)
assistant_message = response.choices[0].message.content
print(assistant_message)
Tool Call
Let's solve some equations thanks to our simple Python calculator tool.
import json
from datetime import datetime, timedelta
from openai import OpenAI
from huggingface_hub import hf_hub_download
# Modify OpenAI's API key and API base to use vLLM's API server.
openai_api_key = "EMPTY"
openai_api_base = "http://localhost:8000/v1"
TEMP = 0.1
client = OpenAI(
api_key=openai_api_key,
base_url=openai_api_base,
)
models = client.models.list()
model = models.data[0].id
def load_system_prompt(repo_id: str, filename: str) -> str:
file_path = hf_hub_download(repo_id=repo_id, filename=filename)
with open(file_path, "r") as file:
system_prompt = file.read()
today = datetime.today().strftime("%Y-%m-%d")
yesterday = (datetime.today() - timedelta(days=1)).strftime("%Y-%m-%d")
model_name = repo_id.split("/")[-1]
return system_prompt.format(name=model_name, today=today, yesterday=yesterday)
SYSTEM_PROMPT = load_system_prompt(model, "SYSTEM_PROMPT.txt")
image_url = "https://math-coaching.com/img/fiche/46/expressions-mathematiques.jpg"
def my_calculator(expression: str) -> str:
return str(eval(expression))
tools = [
{
"type": "function",
"function": {
"name": "my_calculator",
"description": "A calculator that can evaluate a mathematical expression.",
"parameters": {
"type": "object",
"properties": {
"expression": {
"type": "string",
"description": "The mathematical expression to evaluate.",
},
},
"required": ["expression"],
},
},
},
{
"type": "function",
"function": {
"name": "rewrite",
"description": "Rewrite a given text for improved clarity",
"parameters": {
"type": "object",
"properties": {
"text": {
"type": "string",
"description": "The input text to rewrite",
}
},
},
},
},
]
messages = [
{"role": "system", "content": SYSTEM_PROMPT},
{
"role": "user",
"content": [
{
"type": "text",
"text": "Thanks to your calculator, compute the results for the equations that involve numbers displayed in the image.",
},
{
"type": "image_url",
"image_url": {
"url": image_url,
},
},
],
},
]
response = client.chat.completions.create(
model=model,
messages=messages,
temperature=TEMP,
tools=tools,
tool_choice="auto",
reasoning_effort="none",
)
tool_calls = response.choices[0].message.tool_calls
results = []
for tool_call in tool_calls:
function_name = tool_call.function.name
function_args = tool_call.function.arguments
if function_name == "my_calculator":
result = my_calculator(**json.loads(function_args))
results.append(result)
messages.append({"role": "assistant", "tool_calls": tool_calls})
for tool_call, result in zip(tool_calls, results):
messages.append(
{
"role": "tool",
"tool_call_id": tool_call.id,
"name": tool_call.function.name,
"content": result,
}
)
response = client.chat.completions.create(
model=model,
messages=messages,
temperature=TEMP,
reasoning_effort="none",
)
print(response.choices[0].message.content)
Vision Reasoning
Let's see if the Mistral Small 4 knows when to pick a fight !
from datetime import datetime, timedelta
from openai import OpenAI
from huggingface_hub import hf_hub_download
# Modify OpenAI's API key and API base to use vLLM's API server.
openai_api_key = "EMPTY"
openai_api_base = "http://localhost:8000/v1"
TEMP = 0.1
client = OpenAI(
api_key=openai_api_key,
base_url=openai_api_base,
)
models = client.models.list()
model = models.data[0].id
def load_system_prompt(repo_id: str, filename: str) -> str:
file_path = hf_hub_download(repo_id=repo_id, filename=filename)
with open(file_path, "r") as file:
system_prompt = file.read()
today = datetime.today().strftime("%Y-%m-%d")
yesterday = (datetime.today() - timedelta(days=1)).strftime("%Y-%m-%d")
model_name = repo_id.split("/")[-1]
return system_prompt.format(name=model_name, today=today, yesterday=yesterday)
SYSTEM_PROMPT = load_system_prompt(model, "SYSTEM_PROMPT.txt")
image_url = "https://static.wikia.nocookie.net/essentialsdocs/images/7/70/Battle.png/revision/latest?cb=20220523172438"
messages = [
{"role": "system", "content": SYSTEM_PROMPT},
{
"role": "user",
"content": [
{
"type": "text",
"text": "What action do you think I should take in this situation? List all the possible actions and explain why you think they are good or bad.",
},
{"type": "image_url", "image_url": {"url": image_url}},
],
},
]
response = client.chat.completions.create(
model=model,
messages=messages,
temperature=TEMP,
reasoning_effort="high",
)
print(response.choices[0].message.content)
Note: Current implementation of Transformers does not support FP8.
Weights have been stored in FP8 and updates to load them in this format are expected, in the meantime we provide BF16 quantization snippets to ease usage. As soon as support is added, we will update the following code snippet.
Python Inference Snippet
from pathlib import Path
import torch
from huggingface_hub import snapshot_download
from safetensors.torch import load_file
from tqdm import tqdm
from transformers import AutoConfig, AutoProcessor, Mistral3ForConditionalGeneration
def _descale_fp8_to_bf16(tensor: torch.Tensor, scale_inv: torch.Tensor) -> torch.Tensor:
return (tensor.to(torch.bfloat16) * scale_inv.to(torch.bfloat16)).to(torch.bfloat16)
def _resolve_model_dir(model_id: str) -> Path:
local = Path(model_id)
if local.is_dir():
return local
return Path(snapshot_download(model_id, allow_patterns=["model*.safetensors"]))
def load_and_dequantize_state_dict(model_id: str) -> dict[str, torch.Tensor]:
model_dir = _resolve_model_dir(model_id)
shards = sorted(model_dir.glob("model*.safetensors"))
full_state_dict: dict[str, torch.Tensor] = {}
for shard in tqdm(shards, desc="Loading safetensors shards"):
full_state_dict.update(load_file(str(shard)))
scale_suffixes = ("weight_scale_inv", "gate_up_proj_scale_inv", "down_proj_scale_inv", "up_proj_scale_inv")
activation_scale_suffixes = ("activation_scale", "gate_up_proj_activation_scale", "down_proj_activation_scale")
keys_to_remove: set[str] = set()
all_keys = list(full_state_dict.keys())
for key in tqdm(all_keys, desc="Dequantizing FP8 weights to BF16"):
if any(key.endswith(s) for s in scale_suffixes + activation_scale_suffixes):
continue
for scale_suffix in scale_suffixes:
if scale_suffix == "weight_scale_inv":
if not key.endswith(".weight"):
continue
scale_key = key.rsplit(".weight", 1)[0] + ".weight_scale_inv"
else:
proj_name = scale_suffix.replace("_scale_inv", "")
if not key.endswith(f".{proj_name}"):
continue
scale_key = key + "_scale_inv"
if scale_key in full_state_dict:
full_state_dict[key] = _descale_fp8_to_bf16(full_state_dict[key], full_state_dict[scale_key])
keys_to_remove.add(scale_key)
for key in full_state_dict:
if any(key.endswith(s) for s in activation_scale_suffixes):
keys_to_remove.add(key)
for key in tqdm(keys_to_remove, desc="Removing scale keys"):
del full_state_dict[key]
return full_state_dict
def load_config_without_quantization(model_id: str) -> AutoConfig:
config = AutoConfig.from_pretrained(model_id)
if hasattr(config, "quantization_config"):
del config.quantization_config
if hasattr(config, "text_config") and hasattr(config.text_config, "quantization_config"):
del config.text_config.quantization_config
return config
model_id = "mistralai/Mistral-Small-4-119B-2603"
config = load_config_without_quantization(model_id)
state_dict = load_and_dequantize_state_dict(model_id)
model = Mistral3ForConditionalGeneration.from_pretrained(
None,
config=config,
state_dict=state_dict,
device_map="auto",
)
processor = AutoProcessor.from_pretrained(model_id)
image_url = "https://static.wikia.nocookie.net/essentialsdocs/images/7/70/Battle.png/revision/latest?cb=20220523172438"
messages = [
{
"role": "user",
"content": [
{
"type": "text",
"text": "What action do you think I should take in this situation? List all the possible actions and explain why you think they are good or bad.",
},
{"type": "image_url", "image_url": {"url": image_url}},
],
},
]
inputs = processor.apply_chat_template(
messages, return_tensors="pt", tokenize=True, return_dict=True, reasoning_effort="high"
)
inputs = inputs.to(model.device)
output = model.generate(
**inputs,
max_new_tokens=1024,
)[0]
# Setting `skip_special_tokens=False` to visualize reasoning trace between [THINK] [/THINK] tags.
decoded_output = processor.decode(output[len(inputs["input_ids"][0]) :], skip_special_tokens=False)
print(decoded_output)
unsloth <huggingface.co/unsloth>Newsgroups: alt.freeweights.mistralSubject: 
mistralai 
mlx-community