Ministral 3 3B

Ministral 3 3B

Mistral · Released Dec 2, 2025 Professional
Intelligence #14 / 590
82.0 Our Score
Speed #57 / 279
156.1 tokens / sec
Input #195 / 592
$0.100 per 1M tokens
Output #142 / 592
$0.100 per 1M tokens
Context
— Not reported

Analysis Summary

Ministral 3 3B is Mistral's smallest model in the Ministral 3 family, priced at just $0.10 per million tokens for both input and output. At this price point it is one of the most cost-efficient options available, but its intelligence index of 6.8 and coding index of 4.8 reflect the significant capability trade-offs that come with a 3B parameter model. MMLU Pro at 52.4% and GPQA at 35.8% confirm limited performance on complex tasks.

For businesses, this model is best reserved for high-volume, low-complexity tasks: simple classification, keyword extraction, short-form summarisation, or routing logic where cost per call matters more than answer quality. Its agentic index of 24.9 and very low LCR score of 11.7% make it unsuitable for multi-step reasoning or autonomous workflows.

At $0.10/1M tokens it is cheap, but teams should be clear-eyed that the capability ceiling is low. Use it as a fast, cheap first-pass layer in a pipeline, not as a standalone reasoning engine.

Assessed June 30, 2026

Editorial notes

Ministral 3 3B from Mistral is an ultra-cheap small model suited to simple classification or extraction tasks, with limited reasoning and coding capability at this parameter count.

Rankings consider pricing, capabilities, benchmarks, and real-world applicability and are refreshed as new models launch. Feedback?

Performance Profile

Intelligence1.7Technical1.8Value7.3Content2.2
Intelligence 1.7/10
Technical 1.8/10
Content 2.2/10
Value 7.3/10

How Ministral 3 3B compares

Ministral 3 3B ranks #290 of 385 AI models we track for overall intelligence, #134 of 139 for coding, #163 of 293 for agentic tasks. At $0.10 per million input tokens it is cheaper than 67% of comparable models.

Performance Indices

Source: Artificial Analysis

6.8 Intelligence Index
4.8 Coding Index
24.9 Agentic Index
22 Math Index

Benchmark Scores

Intelligence

GPQA Diamond 35.8% Graduate-level scientific reasoning
HLE 5.3% Humanity's Last Exam
MMLU Pro 52.4% Multi-task language understanding
AIME 2025 22% Competition mathematics (2025)
SciCode 14.4% Scientific computing

Technical

LiveCodeBench 24.7% Live coding evaluation
τ²-Bench 24.9% Conversational agent benchmark

Content

IFBench 26.8% Instruction following
LCR 11.7% Long-context reasoning

Benchmark data from Artificial Analysis and Hugging Face

How does Ministral 3 3B stack up?

Compare side-by-side with other professional models.

Compare Models

Model Information

ProviderMistral
Release Date December 2, 2025
Status Active

Pricing

Token Type Cost per 1M tokens Cost per 1K tokens
Input $0.10 $0.000100
Output $0.10 $0.000100

Frequently asked questions about Ministral 3 3B

How much does Ministral 3 3B cost?

Ministral 3 3B costs $0.10 per million input tokens and $0.10 per million output tokens.

Is Ministral 3 3B good for coding?

On our coding benchmark index, Ministral 3 3B ranks #134 of 139 models, placing it in the broader range of the field for code generation and debugging.

Who created Ministral 3 3B?

Ministral 3 3B is developed by Mistral and was released on December 2, 2025.