Ministral 3 14B
Ministral 3 14B is a balanced model in the Ministral 3 family, designed for edge deployment. A powerful, efficient language model with vision capabilities, fine-tuned for instruction tasks. Features multilingual support, strong system prompt adherence, and native function calling. Apache 2.0 licensed.
- Vision
Added Dec 2, 2025
Pricing
Auto routing · per 1M tokens- Input
- $0.10
- Output
- $0.40
Specifications
- Context window
- 262.1K
- Max output
- 32.8K
- Parameters
- 14B
- Avg output (7d)
- 124 tokens
- Longer than 8% of models
Benchmarks
Benchmarks
Sourced from Artificial Analysis.
Intelligence Index
6.0
Coding Index
14.4
Agentic Index
1.1
Agentic work
AutomationBench-AA
Workflow automation with guardrail penalties
0.6%
Better than 10% of models compared
AA-Briefcase
Agentic knowledge work (Elo)
127 Elo
Better than 11% of models compared
GDPval-AA v2
Economically valuable tasks (Elo)
251 Elo
Better than 15% of models compared
Document reasoning
GDP.pdf
Professional PDF reasoning: all-pass rate
1.2%
Better than 9% of models compared
AA-LCR v1.1
Long context reasoning with updated grading
26.3%
Better than 27% of models compared
Reasoning
HLE
Humanity's Last Exam
4.6%
Better than 23% of models compared
IFBench
Instruction-following benchmark
32.0%
Better than 19% of models compared
CritPt
Research-level physics reasoning
0.0%
Coding
Terminal-Bench v4.0
Practical coding and terminal tasks
0.0%
Better than 16% of models compared
SciCode
Python programming for scientific computing
23.8%
Better than 5% of models compared
LiveCodeBench
Contamination-free coding benchmark
35.1%
Better than 42% of models compared
Math
AIME 2025
American Invitational Mathematics Examination 2025
30.0%
Better than 31% of models compared
Knowledge
MMLU-Pro
Professional and academic subject knowledge
69.3%
Better than 34% of models compared
AA-Omniscience Accuracy
Proportion of correctly answered questions
13.6%
AA-Omniscience Hallucination Rate
Rate of incorrect answers among non-correct responses
92.5%
Legacy benchmarks
GPQA Diamond (legacy)
Graduate-level scientific reasoning
57.2%
Better than 31% of models compared
Terminal-Bench Hard (legacy)
Agentic coding and terminal use
4.5%
Better than 29% of models compared
T²-Bench Telecom (legacy)
Conversational AI agents in dual-control scenarios
27.2%
Better than 33% of models compared
AA-LCR (unversioned / legacy)
Long context reasoning evaluation
26.3%
Better than 27% of models compared
GDPval-AA (unversioned / legacy)
Economically valuable tasks
0.0%
Last updated Oct 3, 2026
Artificial AnalysisProviders
Auto routing is available for this model. Explicit provider selection is not available.
Loading provider options…