Ministral Models by MistralAI
5 models2025Up to 32k ctxFrom $0.04/1M input
Capabilities
Vision3 of 5 models
Multimodal3 of 5 models
Structured Outputs2 of 5 models
About
Ministral is a family of 5 AI models by MistralAI, released in 2025.
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
5 in view
Ministral 3 14BCurrent
Use when the workload needs 14B parameters and multimodal inputs.
2025-1214B parametersmultimodal inputs
Mistral Ministral 3 8BCurrent
Use when the workload needs 8B parameters and multimodal inputs.
2025-128B parametersmultimodal inputs
Mistral Ministral 3BCurrent
Use when the workload needs 3B parameters and multimodal inputs.
2025-123B parametersmultimodal inputs
Ministral 14BCurrent
Use when the workload needs 32k context, 14 parameters, and structured outputs.
2025-0632k context14 parametersstructured outputs
Ministral 8BCurrent
Use when the workload needs 32k context, 8 parameters, and structured outputs.
2025-0632k context8 parametersstructured outputs
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Ministral 3 14B | Use when the workload needs 14B parameters and multimodal inputs. | 2025-12 | 14B parametersmultimodal inputs | Current |
| Mistral Ministral 3 8B | Use when the workload needs 8B parameters and multimodal inputs. | 2025-12 | 8B parametersmultimodal inputs | Current |
| Mistral Ministral 3B | Use when the workload needs 3B parameters and multimodal inputs. | 2025-12 | 3B parametersmultimodal inputs | Current |
| Ministral 14B | Use when the workload needs 32k context, 14 parameters, and structured outputs. | 2025-06 | 32k context14 parametersstructured outputs | Current |
| Ministral 8B | Use when the workload needs 32k context, 8 parameters, and structured outputs. | 2025-06 | 32k context8 parametersstructured outputs | Current |
Release Timeline
2 release groups2025-12
3 current
Ministral 3 14B
Current14B parametersmultimodal inputs
Mistral Ministral 3 8B
Current8B parametersmultimodal inputs
Mistral Ministral 3B
Current3B parametersmultimodal inputs
2025-06
2 current
Ministral 14B
Current32k context14 parametersstructured outputs
Ministral 8B
Current32k context8 parametersstructured outputs
Specifications(5 models)
| Model | Released | Context | Parameters | Vision | Multimodal | Structured Outputs |
|---|---|---|---|---|---|---|
| Ministral 3 14B | 2025-12 | — | 14B | Yes | Yes | No |
| Mistral Ministral 3 8B | 2025-12 | — | 8B | Yes | Yes | No |
| Mistral Ministral 3B | 2025-12 | — | 3B | Yes | Yes | No |
| Ministral 14B | 2025-06 | 32k | 14 | No | No | Yes |
| Ministral 8B | 2025-06 | 32k | 8 | No | No | Yes |
Available From(6 providers)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| Mistral Ministral 3B | Microsoft Foundry | $0.04 | $0.04 | Serverless |
| Mistral Ministral 3B | AWS Bedrock | $0.1 | $0.1 | Serverless |
| Ministral 8B | AWS Bedrock | $0.1 | $0.1 | Serverless |
| Mistral Ministral 3B | Vercel AI Gateway | $0.1 | $0.1 | Serverless |
| Mistral Ministral 3B | Mistral AI Studio | $0.1 | $0.1 | Serverless |
| Mistral Ministral 3 8B | AWS Bedrock | $0.15 | $0.15 | Serverless |
| Ministral 8B | Mistral AI Studio | $0.15 | $0.15 | Serverless |
| Ministral 8B | Vercel AI Gateway | $0.15 | $0.15 | Serverless |
| Mistral Ministral 3B | NextBit | $0.15 | $0.15 | Serverless |
| Ministral 3 14B | AWS Bedrock | $0.2 | $0.2 | Serverless |
| Ministral 14B | Mistral AI Studio | $0.2 | $0.2 | Serverless |
| Ministral 14B | Vercel AI Gateway | $0.2 | $0.2 | Serverless |
| Ministral 14B | AWS Bedrock | $0.24 | $0.24 | Serverless |
Frequently Asked Questions
- What is Ministral used for?
- Ministral is used for vision and multimodal work and structured outputs. The family description and listed model capabilities point to those workloads as the best fit.
- How does Ministral compare to Devstral?
- Ministral by MistralAI is strongest where you need vision and multimodal work, while Devstral by MistralAI is the closest related family to check for code. Ministral has 5 listed variants and reaches up to 32k context, while Devstral reaches up to 256k context, so compare the specs and pricing tables before choosing a production model.
- Which Ministral model should I use?
- For the lowest listed input price, start with Mistral Ministral 3B through Microsoft Foundry at $0.04/1M input tokens. For the most capable/latest local choice, evaluate Ministral 3 14B with multimodal inputs.
