M

Mistral Small Creative

Mistral AI
33K
Context
$0.1000
Input /1M
$0.3000
Output /1M
N/A
Max Output
Run API Call Open in Chat

Mistral Small Creative

Model ID for Inference: mistralai/mistral-small-creative

Overview

Property Value
Full Name Mistral: Mistral Small Creative
Model ID mistralai/mistral-small-creative
Provider Model ID labs-mistral-small-creative
Created December 16, 2025
Context Length 32,768 tokens
Input Modalities Text
Output Modalities Text

Description

Mistral Small Creative is an experimental small model designed for:

  • Creative writing - Narrative generation, storytelling, and prose
  • Roleplay and character-driven dialogue - Interactive fiction and character voices
  • General-purpose instruction following - Following complex multi-step instructions
  • Conversational agents - Natural dialogue and conversation flow

This model is optimized for creative tasks while maintaining strong general instruction-following capabilities. It represents Mistral AI's exploration into specialized creative AI assistants.

Pricing

Type Cost per Million Tokens
Input $0.10
Output $0.30

Cost Examples

Use Case Input Tokens Output Tokens Total Cost
Short story (2K in, 4K out) 2,000 4,000 $0.0014
Novel chapter (5K in, 10K out) 5,000 10,000 $0.0035
Extended roleplay session (20K in, 30K out) 20,000 30,000 $0.011

Supported Parameters

Core Parameters

Parameter Supported Notes
temperature Yes Controls randomness (default: 0.3)
top_p Yes Nucleus sampling (default: 0.95)
max_tokens Yes Maximum output length
stop Yes Stop sequences

Tool Use

Parameter Supported Notes
tools Yes Function/tool definitions
tool_choice Yes Control tool selection

Advanced Parameters

Parameter Supported Notes
frequency_penalty Yes Reduce repetition
presence_penalty Yes Encourage topic diversity
seed Yes Reproducible outputs
response_format Yes Structured output format

Reasoning

Parameter Supported Notes
include_reasoning Yes Reasoning configuration available

Performance

Default Parameters

Parameter Default Value
Temperature 0.3
Top P 0.95

Recent Activity (December 16-23, 2025)

Metric Value
Total Prompt Tokens 1.01 trillion
Total Completion Tokens 203.5 billion
Total Requests 598,694
Tool Calls 4,124

Peak Day (December 17, 2025):

  • Requests: 218,369
  • Prompt Tokens: 268.3M
  • Completion Tokens: 51.5M

Best Practices

For Creative Writing

  • Use higher temperature (0.6-0.9) for more creative outputs
  • Provide detailed context and character descriptions
  • Use system prompts to establish tone and style

For Roleplay

  • Define character personalities in the system prompt
  • Use lower temperature (0.3-0.5) for more consistent character behavior
  • Include scenario context for immersive experiences

For Instruction Following

  • Keep the default temperature (0.3) for accuracy
  • Break complex tasks into clear steps
  • Use structured prompts for predictable outputs

From Mistral AI

Model ID Description
mistralai/mistral-nemo Lightweight general-purpose model
mistralai/mistral-small-3.2-24b-instruct-2506 Small instruction-tuned model
mistralai/mistral-small-24b-instruct-2501 Previous small instruction model
mistralai/mistral-tiny Ultra-lightweight model
mistralai/mixtral-8x7b-instruct Mixture of experts model
mistralai/mistral-medium-3.1 Medium-sized general model
mistralai/mistral-large-2411 Large flagship model (Nov 2024)
mistralai/mistral-large-2512 Large flagship model (Dec 2025)
mistralai/codestral-2508 Code-specialized model
mistralai/devstral-2512 Developer-focused model

Providers

Mistral AI (Primary Provider)

Property Value
Provider Mistral AI
Provider Model ID labs-mistral-small-creative
Adapter MistralAdapter
Status Page https://status.mistral.ai/
Data Policy Prompts retained for 30 days max; no training use
BYOK Enabled Yes
Multipart Support Yes

Data Privacy

  • Prompts are retained for a maximum of 30 days
  • No training use of user data
  • Cannot publish or share user content

Usage Example

curl https://api.langmart.ai/v1/chat/completions \
  -H "Authorization: Bearer $LANGMART_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "mistralai/mistral-small-creative",
    "messages": [
      {
        "role": "user",
        "content": "Write the opening paragraph of a fantasy novel about a thief who discovers they can talk to shadows."
      }
    ],
    "temperature": 0.7,
    "max_tokens": 500
  }'

Source: LangMart (https://langmart.ai/model-docs) Last Updated: December 23, 2025