NOUMENOpen app
NOUMEN

Mistral · Open weights

Mistral Saba

Added Jan 1, 2024Open weights

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Use in the app

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.

Input
Text
Output
Text
Discovery tier
Open weights

Pricing

USD per 1M tokens
Input
$0.20
Output
$0.60

Reference prices as published by the model provider. API example

Specifications

Context window
32.8K
Max output
26.2K
Average output (7 days)
63 tokens
Longer than 6% of 468 models

Side by side

Mistral Saba compared

Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Mistral SabaMistralThis modelLlama 4 ScoutMetaRemove
ProviderMistralMeta
Discovery tierOpen weightsOpen weights
AddedJan 1, 2024Sep 5, 2025
Limits
Context window32.77K328KBest
Max output26.21K65.54KBest
Average output63 tokens79 tokens
Pricing, USD per 1M tokens
Input$0.20$0.085Best
Output$0.60$0.46Best
Performance
Request success (24h)100.0%Best49.6%
Benchmark indexes
Intelligence Index6.56.5
Coding Index—8.2
Math Index—14.0
Capabilities
ToolsNoYes
VisionNoYes
Structured outputNoYes
Reasoning effort——

Public evaluations

Benchmarks

Independent evaluation results, with how each score ranks among the models in the Noumen catalog.

  • Mistral Saba
  • Llama 4 Scout

Intelligence Index

  • Mistral Saba6.5
  • Llama 4 Scout6.5

Mistral Saba: Better than 5% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • Mistral Saba: —
    • Llama 4 Scout: 8.2

  • Math Index

    Composite of the math evaluations

    • Mistral Saba: —
    • Llama 4 Scout: 14.0

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • Mistral Saba: —
    • Llama 4 Scout: 0.2%

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • Mistral Saba: —
    • Llama 4 Scout: 0.0%

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • Mistral Saba: —
    • Llama 4 Scout: 0.0%

  • AA-LCRv1.1

    Reasoning across long documents

    • Mistral Saba: —
    • Llama 4 Scout: 27.7%

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • Mistral Saba: 4.3%Top
    • Llama 4 Scout: 3.8%

  • GPQA Diamond

    Graduate-level science questions

    • Mistral Saba: 42.4%
    • Llama 4 Scout: 58.7%Top

  • MMLU-Pro

    Multi-subject knowledge and reasoning

    • Mistral Saba: 61.1%
    • Llama 4 Scout: 75.2%Top

  • IFBench

    Following precise output instructions

    • Mistral Saba: —
    • Llama 4 Scout: 39.5%

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • Mistral Saba: —
    • Llama 4 Scout: 0.0%

  • SciCode

    Scientific computing in Python

    • Mistral Saba: —
    • Llama 4 Scout: 21.3%

  • LiveCodeBench

    Recent competitive programming problems

    • Mistral Saba: —
    • Llama 4 Scout: 29.9%

Math

  • AIME 2025

    Competition mathematics, 2025 problems

    • Mistral Saba: —
    • Llama 4 Scout: 14.0%

  • AIME

    Competition mathematics

    • Mistral Saba: 13.0%
    • Llama 4 Scout: 28.3%Top

  • MATH-500

    Competition math problems

    • Mistral Saba: 67.7%
    • Llama 4 Scout: 84.4%Top

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • Mistral Saba: —
    • Llama 4 Scout: 1.5%

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • Mistral Saba: —
    • Llama 4 Scout: 15.5%

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • Mistral Saba: —
    • Llama 4 Scout: 27.7%

Scores from Artificial Analysis, updated Oct 26, 2026. Artificial Analysis

Live routing data

Performance and reliability

Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.

Request success (24h)
100.0%
Reliable
Hourly request success · last 72 hoursLowest hour 100.0%
  • Healthy (98% or more)
  • Degraded (90–98%)
  • Down (under 90%)
  • No requests

Developers

Use it through the API

Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.

Read the quickstart
curl https://www.noumen.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $NOUMEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "mistralai/mistral-saba",
    "messages": [{ "role": "user", "content": "Explain this stack trace." }]
  }'
import os
from openai import OpenAI

client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
    model="mistralai/mistral-saba",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

Mistral

Other Mistral models, starting with the same family.