Reviews
The short version
- Mistral Small 3.2 24B is a free model from Mistral AI that you can run on your own computer. In our tests it's not one we'd recommend right now: 44 out of 100, #36 of 56.
- It solved 12 of 30 coding jobs and scored 47 on reading documents. On our hardest tasks it scored 21.
- Runs on a 24 GB graphics card or a Mac with 32 GB.
Coding
Our coding test is 30 programming jobs, from small ones like reading time durations or cleaning up messy data to harder ones like a config-file parser or a double-entry ledger. We run each answer against tests the model never sees, and a job only counts if everything passes. Mistral Small 3.2 24B got 12 of 30 right. The best local coders solved 29 of 30.
Reading documents
The second test hands the model things like an expense claim thread, a pay stub or an insurance statement, and asks for specific numbers and dates. Many questions need a bit of math, or noticing a correction further down the email. Mistral Small 3.2 24B scored 47; the best model scored 100.
| Test | Score | Public questions | Secret questions |
|---|---|---|---|
| Coding | 40 | 43 | 39 |
| Reading documents | 47 | 60 | 44 |
| Decisions | 83 | 82 | 83 |
On the 18 hardest tasks (included in the scores above) it scored 21. This number separates the top models.
This model did much better on our public questions than on the secret ones. Sometimes that's luck; it can also mean the model has seen the public questions before.
We tested the full-size version online. The smaller versions you download usually score a little lower.
Can your computer run it?
You can download Mistral Small 3.2 24B in several sizes. Smaller versions fit on more computers and lose a little quality, so pick the biggest one that fits.
| Your computer | Runs it? | Version to download |
|---|---|---|
| Graphics card, 8 GB | No | – |
| Graphics card, 12 GB | No | – |
| Graphics card, 16 GB | No | – |
| Graphics card, 24 GB | Yes | Q6 |
| Graphics card, 32 GB | Yes | Q8 |
| Mac, 16 GB | No | – |
| Mac, 24 GB | No | – |
| Mac, 32 GB | Yes | Q5 |
| Mac, 48 GB | Yes | Q8 |
| Mac, 64 GB | Yes | Q8 |
| Mac, 96 GB | Yes | Q8 |
| Mac, 128 GB | Yes | Q8 |
For normal-length chats. Long documents need more memory. Q4 or Q8? · How we worked this out
How to try it
Search for "Mistral Small 3.2 24B GGUF" in LM Studio or on Hugging Face and pick the version that fits your computer. New to this? Start with our Ollama guide.
Want the strongest all-rounder instead? Qwen3.6 27B scored 97.
Its actual answers
What Mistral Small 3.2 24B wrote for our public questions, word for word.
Coding: INI parser (hard) Didn't work
```python
import re
from collections import defaultdict
def parse_ini(text: str) -> dict[str, dict[str, str]]:
sections = defaultdict(dict)
current_section = None
current_key = None
lines = text.splitlines()
default_section = {}
for line in lines:
stripped_line = line.strip()
if not stripped_line or stripped_line.startswith((';', '#')):
continue
# Check for section header
section_match = re.match(r'^\s*\[([^\]]+)\]\s*$', line)
if section_match:
current_section = section_match.group(1).strip()
if current_section.lower() == 'default':
current_section = None
else:
sections[current_section] = {}
continue
# Check for key-value pair
if '=' in line and ':' in line:
sep = min(line.find('='), line.find(':'))
elif '=' in line:
sep = line.find('=')
elif ':' in line:
sep = line.find(':')
else:
raise ValueError(f"Malformed line: {line}")
key_part = line[:sep].strip().lower()
value_part = line[sep+1:].strip()
# Check for inline comment
inline_comment_match = re.search(r'(?<!\\)[\s]*(;|#).*$', value_part)
if inline_comment_match:
value_part = value_part[:inline_comment_match.start()].strip()
# Check for quoted value
if value_part.startswith('"') and value_part.endswith('"'):
value_part = value_part[1:-1]
else:
value_part = value_part.strip()
# Handle continuation lines
if current_key is not None:
if line.startswith((' ', '\t')):
value_part = sections[current_section][current_key] + '\n' + value_part
sections[current_section][current_key] = value_part
continue
else:
current_key = None
current_key = key_part
sections[current_section][current_key] = value_part
# Process DEFAULT section
if None in sections:
default_section = sections.pop(None)
# Apply interpolation
for section_name, section in sections.items():
section.update(default_section)
for key in list(section.keys()):
section[key] = interpolate_value(section, key, sections)
return sections
def interpolate_value(section: dict[str, str], key: str, all_sections: dict[str, dict[str, str]]) -> str:
value = section[key]
pattern = re.compile(r'\$\{(?:([^:]+):)?([^}]+)\}')
while True:
match = pattern.search(value)
if not match:
break
section_name, inner_key = match.groups()
if section_name is None:
if inner_key not in section:
raise KeyError(f"Missing key: {inner_key}")
replacement = section[inner_key]
else:
if section_name not in all_sections:
raise KeyError(f"Missing section: {section_name}")
if inner_key not in all_sections[section_name]:
raise KeyError(f"Missing key: {inner_key} in section {section_name}")
replacement = all_sections[section_name][inner_key]
value = value[:match.start()] + replacement + value[match.end():]
return value
```Coding: TTL LRU (hard) Didn't work
```python
from collections import OrderedDict
class TTLCache:
def __init__(self, capacity: int, ttl: float, clock):
self.capacity = capacity
self.ttl = ttl
self.clock = clock
self.cache = OrderedDict()
self.stats = {
"hits": 0,
"misses": 0,
"evictions": 0,
"expirations": 0
}
def _clean_expired(self):
now = self.clock()
expired_keys = []
for key, (value, expiry) in list(self.cache.items()):
if now >= expiry:
expired_keys.append(key)
for key in expired_keys:
del self.cache[key]
self.stats["expirations"] += 1
def put(self, key, value):
if self.capacity == 0:
return
now = self.clock()
expiry = now + self.ttl
if key in self.cache:
# Update existing entry
del self.cache[key]
self.cache[key] = (value, expiry)
self.cache.move_to_end(key)
else:
# Insert new entry
if len(self.cache) >= self.capacity:
self._clean_expired()
if len(self.cache) >= self.capacity:
evicted_key, _ = self.cache.popitem(last=False)
self.stats["evictions"] += 1
self.cache[key] = (value, expiry)
self.cache.move_to_end(key)
def get(self, key, default=None):
if self.capacity == 0:
self.stats["misses"] += 1
return default
now = self.clock()
if key in self.cache:
value, expiry = self.cache[key]
if now >= expiry:
del self.cache[key]
self.stats["expirations"] += 1
self.stats["misses"] += 1
return default
else:
self.cache.move_to_end(key)
self.stats["hits"] += 1
return value
else:
self.stats["misses"] += 1
return default
def __len__(self):
now = self.clock()
return sum(1 for _, (_, expiry) in self.cache.items() if now < expiry)
def stats(self):
return self.stats.copy()
```Coding: Parse duration Didn't work
```python
def parse_duration(s: str) -> int:
if not s or s.isspace():
raise ValueError("Empty or blank string")
parts = []
for part in s.split():
if not part:
continue
if 'h' in part:
if part.lower() != 'h':
raise ValueError("Invalid format for hours")
parts.append(('h', 1))
elif 'm' in part:
if part.lower() != 'm':
raise ValueError("Invalid format for minutes")
parts.append(('m', 2))
elif 's' in part:
if part.lower() != 's':
raise ValueError("Invalid format for seconds")
parts.append(('s', 3))
else:
raise ValueError("Unknown unit")
if not parts:
raise ValueError("No valid parts found")
# Check for repeated units
units = [unit for unit, _ in parts]
if len(units) != len(set(units)):
raise ValueError("Repeated units")
# Check order of units
expected_order = ['h', 'm', 's']
for i, (unit, _) in enumerate(parts):
if i < len(expected_order) and unit != expected_order[i]:
raise ValueError("Units out of order")
# Parse numbers
total_seconds = 0
for i, (unit, multiplier) in enumerate(parts):
try:
num = int(part.split(unit)[0])
except ValueError:
raise ValueError("Invalid number format")
if num < 0:
raise ValueError("Negative number")
if unit == 'h':
total_seconds += num * 3600
elif unit == 'm':
total_seconds += num * 60
elif unit == 's':
total_seconds += num
return total_seconds
```Coding: Summarize ranges Worked
```python
def summarize_ranges(nums: list[int]) -> str:
if not nums:
return ""
unique_sorted = sorted(set(nums))
result = []
start = unique_sorted[0]
for i in range(1, len(unique_sorted)):
if unique_sorted[i] != unique_sorted[i - 1] + 1:
if unique_sorted[i - 1] - start >= 2:
result.append(f"{start}..{unique_sorted[i - 1]}")
else:
for num in range(start, unique_sorted[i - 1] + 1):
result.append(str(num))
start = unique_sorted[i]
# Handle the last range
if unique_sorted[-1] - start >= 2:
result.append(f"{start}..{unique_sorted[-1]}")
else:
for num in range(start, unique_sorted[-1] + 1):
result.append(str(num))
return ",".join(result)
```Coding: Top customers Worked
```python
def top_customers(rows: list[dict], n: int) -> list[tuple[str, float]]:
customer_totals = {}
for row in rows:
customer = row.get("customer", "").strip()
amount_str = row.get("amount", "")
if not customer:
continue
# Parse the amount
try:
amount = 0.0
# Handle negative amounts
if amount_str.startswith("-"):
amount_str = amount_str[1:]
negative = True
else:
negative = False
# Remove any non-digit characters except decimal point
cleaned = ''.join(c for c in amount_str if c.isdigit() or c == '.')
if not cleaned:
continue
# Parse the cleaned amount
amount = float(cleaned)
if negative:
amount = -amount
except (ValueError, TypeError):
continue
# Get the first valid customer name (case-insensitive but stored as-is)
if customer.lower() not in customer_totals:
customer_totals[customer.lower()] = (customer, 0.0)
# Add to the total
name, total = customer_totals[customer.lower()]
customer_totals[customer.lower()] = (name, total + amount)
# Prepare the list of (name, total) tuples
customer_list = [(name, round(total, 2)) for name, total in customer_totals.values()]
# Sort by total descending, then name ascending
customer_list.sort(key=lambda x: (-x[1], x[0]))
# Return the top n customers
return customer_list[:n]
```Coding: Slugify Didn't work
```python
import unicodedata
import re
def slugify(title: str, max_len: int = 50) -> str:
# Step 1: Transliterate accented characters to ASCII and drop non-ASCII
normalized = unicodedata.normalize('NFKD', title)
ascii_chars = []
for c in normalized:
if unicodedata.category(c) == 'Mn':
continue
if ord(c) < 128:
ascii_chars.append(c)
ascii_str = ''.join(ascii_chars)
# Step 2: Lowercase
lower_str = ascii_str.lower()
# Step 3: Replace non-alphanumeric runs with "-"
slug = re.sub(r'[^a-z0-9]+', '-', lower_str)
slug = slug.strip('-')
# Step 4: Truncate to max_len if needed
if len(slug) > max_len:
words = slug.split('-')
truncated = []
remaining = max_len
for word in words:
if remaining >= len(word) + 1: # +1 for the potential "-"
truncated.append(word)
remaining -= len(word) + 1
else:
break
if not truncated:
# First word is too long, hard-cut it
truncated = [slug[:max_len]]
slug = '-'.join(truncated)
return slug
```Coding: Token bucket Worked
```python
import time
from typing import Callable
class TokenBucket:
def __init__(self, capacity: float, refill_per_sec: float, clock: Callable[[], float]):
if capacity <= 0 or refill_per_sec < 0:
raise ValueError("Capacity must be positive and refill rate must be non-negative")
self._capacity = capacity
self._refill_rate = refill_per_sec
self._clock = clock
self._last_refill_time = clock()
self._tokens = capacity
def _refill(self) -> None:
now = self._clock()
elapsed = now - self._last_refill_time
if elapsed > 0:
refill_amount = elapsed * self._refill_rate
self._tokens = min(self._tokens + refill_amount, self._capacity)
self._last_refill_time = now
def allow(self, cost: float = 1) -> bool:
self._refill()
if self._tokens >= cost:
self._tokens -= cost
return True
return False
@property
def tokens(self) -> float:
self._refill()
return self._tokens
```Decisions: Refund window (hard) 50% right
```json
{
"outcome": {
"full_refund": 0.0,
"partial_refund": 0.9,
"store_credit": 0.1,
"nothing": 0.0
},
"defective": {
"true": 0.0,
"false": 1.0
}
}
```Decisions: Incident matrix (hard) 0% right
```json
{
"severity": {
"0": 0.0,
"1": 0.1,
"2": 0.9,
"3": 0.0
},
"page": {
"true": 0.9,
"false": 0.1
}
}
```Decisions: Tool followup (hard) 100% right
```json
{
"tool": {
"calendar": 1.0,
"web_search": 0.0,
"calculator": 0.0,
"email": 0.0,
"none": 0.0
},
"confirm": {
"true": 1.0,
"false": 0.0
}
}
```Decisions: Legit security alert (hard) 100% right
```json
{
"phishing": {
"true": 0.05,
"false": 0.95
},
"action_needed": {
"true": 0.1,
"false": 0.9
}
}
```Decisions: Meeting slot (hard) 0% right
```json
{
"slot": {
"A": 0,
"B": 0.25,
"C": 0.75,
"D": 0
},
"raj_last": {
"true": 0,
"false": 1
}
}
```Decisions: Review mixed (hard) 100% right
```json
{
"hardware": {
"true": 1.0,
"false": 0.0
},
"support": {
"true": 1.0,
"false": 0.0
}
}
```Decisions: Support checkout down 100% right
```json
{
"department": {
"billing": 0.1,
"technical": 0.8,
"account": 0.05,
"shipping": 0.02,
"sales": 0.03
},
"urgency": {
"0": 0,
"1": 0,
"2": 0.1,
"3": 0.9
},
"outage": {
"true": 0.9,
"false": 0.1
}
}
```Decisions: Refund wrong plan 100% right
```json
{
"department": {
"billing": 0.9,
"technical": 0.05,
"account": 0.05,
"shipping": 0,
"sales": 0
},
"refund": {
"true": 1,
"false": 0
},
"tone": {
"frustrated": 0.1,
"calm": 0.9
}
}
```Decisions: Moderation doxxing 100% right
```json
{
"policy": {
"none": 0.0,
"harassment": 0.9,
"hate": 0.0,
"spam": 0.0,
"self_harm": 0.1
},
"personal_info": {
"true": 0.9,
"false": 0.1
}
}
```Decisions: Route calendar 100% right
```json
{
"tool": {
"web_search": 0.0,
"calculator": 0.0,
"calendar": 0.95,
"email": 0.05,
"none": 0.0
},
"confirm": {
"true": 0.8,
"false": 0.2
}
}
```Decisions: Doc invoice missing due 100% right
```json
{
"doc_type": {
"invoice": 0.95,
"resume": 0.01,
"contract": 0.02,
"bank_statement": 0.01,
"other": 0.01
},
"missing_due_date": {
"true": 0.9,
"false": 0.1
}
}
```Decisions: Phishing paypal 100% right
```json
{
"phishing": {
"true": 1.0,
"false": 0.0
},
"risk": {
"0": 0.0,
"1": 0.0,
"2": 0.0,
"3": 1.0
}
}
```Decisions: Pii ssn email 100% right
```json
{
"data_kind": {
"none": 0.0,
"contact": 0.2,
"financial": 0.1,
"government_id": 0.7,
"health": 0.0
},
"sensitive": {
"true": 0.7,
"false": 0.3
}
}
```Decisions: Review mixed 100% right
```json
{
"sentiment": {
"positive": 0.0,
"neutral": 0.0,
"negative": 1.0
},
"defect": {
"true": 1.0,
"false": 0.0
},
"recommend": {
"true": 0.0,
"false": 1.0
}
}
```Documents: Saas escalator (hard) 30% right
```json
{
"year2_price_per_seat_month": 46.47,
"year3_price_per_seat_month": 46.47,
"year1_invoice": 64800.00,
"year2_invoice": 67764.00,
"addon_months_billed": 6,
"addon_invoice": 11808.00,
"year3_invoice": 67764.00,
"year3_discount_percent": 15,
"total_contract_value": 211336.00,
"contract_end_date": "2027-02-28"
}
```Documents: Expense thread 75% right
```json
{
"employee_id": "EMP-20417",
"destination_city": "Lisbon",
"trip_start": "2025-02-24",
"trip_end": "2025-02-27",
"approved_items": [
{
"date": "2025-02-24",
"category": "airfare",
"amount_usd": 1184.60
},
{
"date": "2025-02-24",
"category": "ground_transport",
"amount_usd": 38.88
},
{
"date": "2025-02-25",
"category": "meals",
"amount_usd": 228.19
},
{
"date": "2025-02-26",
"category": "lodging",
"amount_usd": 432.00
},
{
"date": "2025-02-27",
"category": "ground_transport",
"amount_usd": 44.77
}
],
"rejected_item_count": 1,
"per_diem_days": 2,
"per_diem_usd": 130.00,
"total_reimbursable_usd": 2068.44,
"approver_email": "priya.raman@corvane.com"
}
```Documents: Lease amendment 92% right
```json
{
"tenants": ["Marcus Lin", "Sofia Lin"],
"landlord": "Ridgeline Property Group LLC",
"zip": "97205",
"lease_end": "2025-11-30",
"original_monthly_rent": 2150.00,
"monthly_rent_from_2025_06_01": 2236.00,
"late_fee_from_2025_06_01": 107.80,
"security_deposit": 2150.00,
"total_pet_deposits": 800.00,
"total_monthly_payment_july_2025": 2306.00,
"move_in_payment": 4700.00
}
```Documents: Ticket SLA 82% right
```json
{
"ticket_id": "48213",
"account_id": "ACC-7731",
"open_issue": "inventory_sync",
"resolved_issues": ["billing_address"],
"affected_orders": ["SO-99812", "SO-99820", "SO-99827"],
"priority": "P2",
"sla_due_local": "2025-09-12T23:30",
"sla_due_utc": "2025-09-13T04:30:00Z",
"reissued_invoice": "INV-2025-0812"
}
```Documents: Sales footnotes 22% right
```json
{
"q3_total_usd": 15346,
"q2_total_usd": 14464,
"q2_central_originally_reported_usd": 3047,
"q2_to_q3_change_pct": 5.4,
"top_region_q3": "East",
"fastest_growing_region_q1_to_q3": "International",
"regions_declining_q2_to_q3": ["Central"],
"international_q3_organic_usd": 1731,
"west_excluding_mountain_q3_usd": 4201
}
```Size: 24B parameters. First tested OCT 10.
Comments
Sign in with GitHub to comment. Spam and abuse are hidden automatically.