Mistral Saba
Neural Network
Mistral Saba is a 24B parameter Mistral AI language model focused on the regional context of the Middle East and South Asia.
Specifications
- Input
- text, documents
- Output
- text
- Context size
- 32 768 tokens
- Max answer length
- 26 214 tokens
- Capabilities
- function calling
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
import OpenAI from 'openai';
const openai = new OpenAI({
apiKey: '<your bothub access token>',
baseURL: 'https://openai.bothub.chat/v1'
});
// Sync - Text generation
async function main() {
const chatCompletion = await openai.chat.completions.create({
messages: [{ role: 'user', content: 'Say this is a test' }],
model: 'mistral-saba',
});
}
// Async - Text generation
async function main() {
const stream = await openai.chat.completions.create({
messages: [{ role: 'user', content: 'Say this is a test' }],
model: 'mistral-saba',
stream: true
});
for await (const chunk of stream) {
const part: string | null = chunk.choices[0].delta?.content ?? null;
}
}
main();
How it works Mistral Saba?
Mistral Saba is a 24B parameter language model from Mistral AI, trained on selected regional data and designed for accurate, context-aware responses for the Middle East and South Asia. Its compact size ensures efficiency: the model balances response quality with resource usage, which is noticeable with high-volume requests. In BotHub, it connects instantly without VPN or foreign cards: pay with a Russian card in rubles only for the tokens you use, which never expire. With over 250 models in one window, you can easily compare Mistral Saba's answers with others and switch without rewriting your integration: the API is unified and OpenAI-compatible. Chats are encrypted, and for companies, we offer contracts, invoices, EDI, and an admin panel with limits. Use cases include: support and chatbots with regional specifics, content preparation and adaptation for Middle East and South Asian markets, document analysis and summarization, regional analytical reviews, and embedding the model into multimodal services via API.Frequently asked questions about Mistral Saba
You can use generated results for commercial purposes. You own all rights to the content you create. The only restriction: make sure your prompt does not include copyrighted third-party material. You are responsible for respecting the rights to any input data.
Mistral Saba works with text and documents: it parses files, summarizes, answers questions about content, writes and edits code, and helps with analytics and correspondence. A context window of up to 32,000 tokens allows you to work with long instructions, reports, and entire project fragments.
The output limit is 26,214 tokens — enough for a lengthy article, detailed instruction, comprehensive document analysis, or a large code module. The total dialogue context is limited to 32,000 tokens, including your request and the model's response.
A separate step-by-step reasoning mode is not specified in our data. In practice, you can simply ask the model to write out the solution step-by-step, break the task into parts, or explain the logic — this technique works even without a special reasoning mode.
Yes, the model supports function calling and structured JSON output. This is useful for agents, parsing documents into fields, and connecting external services and databases. BotHub provides a unified OpenAI-compatible API, so switching models won't require rewriting your integration.
The model accepts text and documents, including natively — you can upload a file and ask questions about it. For images, audio, and video, switch to a multimodal model in BotHub in the same window, without a separate subscription or VPN.
Mistral Saba was created with a focus on Middle Eastern and South Asian languages, so quality varies by language. Test the model on your task: in BotHub, you pay per token, and you can compare answers with other models in the same interface.