modelgrep
M

Mistral: Ministral 3 3B 2512

mistralai/ministral-3b-2512

162nd smartest of 178Cheaper than 83% of paidToolsJSONVision
Use via OpenRouter ↗
Intelligence
11.2
162nd of 178
Design Elo
1072
Website
Speed
54
153rd fastest
Latency
247ms
first token
Input price
$0.100
52nd cheapest
Context
131K

How it compares

Smarter than9%
of all ranked models
Faster than48%
of all ranked models
Cheaper than83%
of all ranked models

Overview

The smallest model in the Ministral 3 family, Ministral 3 3B is a powerful, efficient tiny language model with vision capabilities.

Benchmarks

independent · via OpenRouter
Artificial Analysis10th percentile
Intelligence Index
11.2
Coding Index
4.8
Agentic Index
11.4
GPQA Diamond
36%
Humanity's Last Exam
5%
SciCode
14%
Tau²-Bench (agentic)
25%
Design Arena · Elo1,253 tournaments
Website
1072
codecategories
1062
3D
1043
Game Dev
1018

Providers & pricing (2)

ProviderIn $/MOut $/MUptime
Mistral$0.100$0.10099.9%
NextBitfp8$0.150$0.150100%

Specifications

Context window131K
Max output
Knowledge cutoff
Input modalitiestext, image
Output modalitiestext
Prompt caching
Cache read price$0.010/M
ModeratedNo

Ministral 3 3B 2512 FAQ

How much does Ministral 3 3B 2512 cost?

Ministral 3 3B 2512 costs $0.100 per million input tokens and $0.100 per million output tokens via OpenRouter, making it 52nd cheapest of 298 paid models.

How smart is Ministral 3 3B 2512?

Ministral 3 3B 2512 scores 11.2 on the Artificial Analysis Intelligence Index, ranking 162nd of 178 benchmarked models, with a GPQA Diamond score of 36%.

How fast is Ministral 3 3B 2512?

Ministral 3 3B 2512 generates around 54 tokens per second with 247ms time-to-first-token (p50), the 153rd fastest tracked model.

What is Ministral 3 3B 2512's context window?

Ministral 3 3B 2512 supports a 131K-token context window. It accepts text, image input.

Compare head-to-head