Mistral AIOpen weightsAuto-detected

Ministral 3 3B

Name: Ministral 3 3B benchmark results
Creator: BenchLeader

Best configuration ranks #559 of 610 on the BenchLeader Index at 37.6 ±3.6. Released 2 Dec 2025.

Blended price: $0.100/M; $0.100 in · $0.100 out
Output speed: 184 tok/s
First answer: 0.66 s; first token 0.66 s
Context: 131k

How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.

Overall index38
Agents & tools34
Knowledge34
Instruction following31
Multimodal22
Long context34
Composite35

Benchmark results

One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.

Reasoning

Benchmark	default	Source
GPQA Diamond (AA)not in index	35.8%#463	Artificial Analysis
Humanity's Last Exam (AA)not in index	5.4%#354	Artificial Analysis

Coding

Benchmark	default	Source
SciCode (AA)not in index	15.3%#161	Artificial Analysis

Agents & tools

Benchmark	default	Source
Terminal-Bench Hard	0.0%#345	Artificial Analysis
τ²-Bench Telecom (AA)not in index	24.9%#285	Artificial Analysis

Knowledge

Benchmark	default	Source
AA-Omniscience	-64#414	Artificial Analysis

Instruction following

Benchmark	default	Source
IFBench	26.8%#364	Artificial Analysis

Multimodal

Benchmark	default	Source
MMMU-Pro	38.1%#232	Artificial Analysis

Long context

Benchmark	default	Source
AA-LCR	17.0%#373	Artificial Analysis

Composite

Benchmark	default	Source
AA Intelligence Index	4.8#526	Artificial Analysis

What a task costs

Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache. Reasoning tokens are not modelled.

Workload	Tokens in / out	Cost	With caching	Time
Chat reply	400 / 300	$0.0001	–	2.3 s
Summarise a 30-page report	12,000 / 600	$0.0013	–	3.9 s
Code edit	6,000 / 1,500	$0.0008	–	8.8 s
Agentic coding session	60,000 / 4,000	$0.0064	–	22.4 s
Structured extraction	2,000 / 200	$0.0002	–	1.7 s