Saba
VerifiedOpen-weight multimodal model for seamless text and file handling.
About Saba
Saba combines multimodal capabilities with an extensive context window of 32768 tokens. Developed by Mistral as an open-weight release, its design emphasizes compatibility with both textual data and file-based inputs. This architecture allows efficient integration into diverse systems without proprietary restrictions.
Its strengths lie in managing mixed modalities while maintaining coherence over long contexts. Users benefit from the transparency of open weights for customization and auditing. Typical applications include document analysis, content generation from files, and multimodal reasoning workflows.
Capabilities
How Saba compares
Saba (striped bar) vs other multimodal on intelligence, speed and price.
Price
USD per 1M output tokens · Lower is better · Saba ranks #30 of 139
Sources: Artificial Analysis (intelligence, speed) · OpenRouter (price).
Best for
Long Document Analysis with Images
Saba processes extended reports and papers containing charts or diagrams, delivering summaries and answers drawn from up to 32768 tokens of mixed content.
Extended Multimodal Conversations
It maintains coherent dialogue across lengthy exchanges that incorporate both text prompts and visual references without losing context.
Visual-Text Reasoning Workflows
Saba integrates visual inputs with detailed textual instructions for tasks such as interpreting infographics or annotating image sequences.
Strengths & limitations
Strengths
- +Effective handling of text and file inputs
- +Solid reasoning within context limits
- +Mistral's efficient model design
- +Versatile for document-based tasks
Limitations
- –32k token context restricts very long inputs
- –File modality support may need specific formats
- –No native audio or video processing
Cost calculator
Estimate what Saba would cost for your usage.
Based on Saba's $0.20/1M input · $0.60/1M output. Estimate only — actual cost varies by provider and caching.
Quick start
OpenRouter's API is OpenAI-compatible — most SDKs work by just swapping the base URL. Only the model slug changes between models.
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://openrouter.ai/api/v1",
apiKey: process.env.OPENROUTER_API_KEY,
});
const completion = await client.chat.completions.create({
model: "mistralai/mistral-saba",
messages: [{ role: "user", content: "Hello!" }],
});
console.log(completion.choices[0].message.content);Model slug: mistralai/mistral-saba
Editor's verdict
Saba is Mistral's open-weight multimodal with a 33K-token context window.
At $0.60 per 1M output tokens, it is very cost-efficient for its class.
As an open-weight model you can self-host it or call it through a hosted API.
Best suited to effective handling of text and file inputs and solid reasoning within context limits.
Frequently asked questions
Saba provides a context window of 32768 tokens.
User reviews
Real, verified reviews from the community shape this model's rating.
Loading reviews…
Similar models
Other multimodal worth comparing.