mistral-cost-tuning
Installation
SKILL.md
Mistral AI Cost Tuning
Overview
Optimize Mistral AI costs through model selection, token management, caching, batch inference, and budget monitoring. Mistral offers the best price-performance in the market with models from $0.1/M tokens (Ministral/Small) to $0.5/M tokens (Large).
Prerequisites
- Access to Mistral AI console for usage data
- Understanding of current usage patterns
- Database for usage tracking (optional)
Pricing Reference (as of 2025)
| Model | Input $/M tokens | Output $/M tokens | Best For |
|---|---|---|---|
ministral-latest (3B) |
$0.10 | $0.10 | Simple tasks, edge |
mistral-small-latest |
$0.10 | $0.30 | General purpose, fast |
codestral-latest |
$0.30 | $0.90 | Code generation |
mistral-large-latest |
$0.50 | $1.50 | Complex reasoning |
pixtral-large-latest |
$2.00 | $6.00 | Vision + text |
mistral-embed |
$0.10 | — | Embeddings |
Related skills