Mistral AI
Mistral: Saba
mistralai/mistral-saba
Mistral Saba is a 24B-parameter language model specifically designed for the Middle East and South Asia, delivering accurate and contextually relevant responses while maintaining efficient performance. Trained on curated regional datasets, it supports multiple Indian-origin languages—including Tamil and Malayalam—alongside Arabic. This makes it a versatile option for a range of regional and multilingual applications. Read more at the blog post here
Input
$0.2
/ 1M tokens
Output
$0.6
/ 1M tokens
Context window
32.77K
tokens
Limits
Context
32,768
tokens
Max output
26,214
80% of the context window
Model facts
Author
Mistral AI
Tokenizer
Mistral
Modality
Text + File -> Text
Knowledge cutoff
2024-09-30
Created
Feb 17, 2025
Default generation parameters
temperature
0.3
Publication status
Hugging Face: unavailableLicense: not publishedParameter count: not published
Modalities
Input · TextInput · FileOutput · Text
Generation controls
Frequency penaltyMax tokensPresence penaltyResponse formatSeedStopStructured outputsTemperatureTool choiceToolsTop p
Capability scores
6.5
Intelligence index
Better than 6% of models compared