Mistral AIOpen weightsAuto-detected

Ministral 3 8B

Name: Ministral 3 8B benchmark results
Creator: BenchLeader

Best configuration ranks #526 of 610 on the BenchLeader Index at 39.5 ±2.5. Released 2 Dec 2025.

Blended price: $0.150/M; $0.150 in · $0.150 out
Output speed: 87 tok/s
First answer: 0.77 s; first token 0.77 s
Context: 262k

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

Overall index40
Agents & tools38
Knowledge31
Instruction following33
Multimodal30
Long context38
Composite36

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	47.1%#419	Artificial Analysis
Humanity's Last Exam (AA)not in index	4.3%#434	Artificial Analysis

Coding

Benchmark	default	Source
SciCode (AA)not in index	20.7%#159	Artificial Analysis

Agents & tools

Benchmark	default	Source
Terminal-Bench Hard	4.5%#270	Artificial Analysis
τ²-Bench Telecom (AA)not in index	26.6%#271	Artificial Analysis

Knowledge

Benchmark	default	Source
AA-Omniscience	-69.0#431	Artificial Analysis

Instruction following

Benchmark	default	Source
IFBench	29.1%#349	Artificial Analysis

Multimodal

Benchmark	default	Source
MMMU-Pro	46.0%#217	Artificial Analysis

Long context

Benchmark	default	Source
AA-LCR	25.7%#338	Artificial Analysis

Composite

Benchmark	default	Source
AA Intelligence Index	5.5#490	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	–	4.2 s
Summarise a 30-page report	12,000 / 600	$0.0019	–	7.7 s
Code edit	6,000 / 1,500	$0.0011	–	18.0 s
Agentic coding session	60,000 / 4,000	$0.0096	–	46.7 s
Structured extraction	2,000 / 200	$0.0003	–	3.1 s