Mistral Small 3.2 24B Instruct
Multimodal
Zero-eval
#1HumanEval Plus
#1IF
#1MBPP Plus
+1 more
by Mistral AI
About
Mistral Small 3.2 24B Instruct is a multimodal language model developed by Mistral AI. It achieves strong performance with an average score of 69.8% across 15 benchmarks. It excels particularly in DocVQA (94.9%), AI2D (92.9%), HumanEval Plus (92.9%). The model shows particular specialization in code tasks with an average performance of 85.6%. As a multimodal model, it can process and understand text, images, and other input formats seamlessly. It's licensed for commercial use, making it suitable for enterprise applications. Released in 2025, it represents Mistral AI's latest advancement in AI technology.
Timeline
AnnouncedJun 20, 2025
ReleasedJun 20, 2025
Knowledge CutoffOct 1, 2023
Specifications
Capabilities
Multimodal
License & Family
License
Apache 2.0
Base ModelMistral Small 3.1 24B Base
Benchmark Performance Overview
Performance metrics and category breakdown
Overall Performance
15 benchmarks
Average Score
69.8%
Best Score
94.9%
High Performers (80%+)
6Top Categories
code
85.6%
vision
78.7%
math
68.3%
general
64.6%
Benchmark Performance
Top benchmark scores with normalized values (0-100%)
Ranking Across Benchmarks
Position relative to other models on each benchmark
DocVQA
Rank #5 of 26
#2Claude 3.5 Sonnet
95.2%
#3Qwen2.5-Omni-7B
95.2%
#4Qwen2.5 VL 7B Instruct
95.7%
#5Mistral Small 3.2 24B Instruct
94.9%
#6Qwen2.5 VL 32B Instruct
94.8%
#7Llama 4 Maverick
94.4%
#8Llama 4 Scout
94.4%
AI2D
Rank #4 of 17
#1Pixtral Large
93.8%
#2GPT-4o
94.2%
#3Claude 3.5 Sonnet
94.7%
#4Mistral Small 3.2 24B Instruct
92.9%
#5Llama 3.2 90B Instruct
92.3%
#6Llama 3.2 11B Instruct
91.1%
#7Qwen2.5 VL 72B Instruct
88.4%
HumanEval Plus
Rank #1 of 1
#1Mistral Small 3.2 24B Instruct
92.9%
ChartQA
Rank #8 of 24
#5Pixtral Large
88.1%
#6Qwen2-VL-72B-Instruct
88.3%
#7Llama 4 Scout
88.8%
#8Mistral Small 3.2 24B Instruct
87.4%
#9Qwen2.5 VL 7B Instruct
87.3%
#10Nova Lite
86.8%
#11DeepSeek VL2
86.0%
IF
Rank #1 of 1
#1Mistral Small 3.2 24B Instruct
84.8%
All Benchmark Results for Mistral Small 3.2 24B Instruct
Complete list of benchmark scores with detailed information
DocVQA DocVQA benchmark | vision | multimodal | 0.95 | 94.9% | Self-reported |
AI2D AI2D benchmark | general | text | 0.93 | 92.9% | Self-reported |
HumanEval Plus HumanEval Plus benchmark | code | text | 0.93 | 92.9% | Self-reported |
ChartQA ChartQA benchmark | general | multimodal | 0.87 | 87.4% | Self-reported |
IF IF benchmark | general | text | 0.85 | 84.8% | Self-reported |
MMLU MMLU benchmark | general | text | 0.81 | 80.5% | Self-reported |
MBPP Plus MBPP Plus benchmark | code | text | 0.78 | 78.3% | Self-reported |
MATH MATH benchmark | math | text | 0.69 | 69.4% | Self-reported |
MMLU-Pro MMLU-Pro benchmark | general | text | 0.69 | 69.1% | Self-reported |
MathVista MathVista benchmark | math | text | 0.67 | 67.1% | Self-reported |
Showing 1 to 10 of 15 benchmarks
Resources