NOUMENOpen app
NOUMEN

Mistral · Open weights

Mistral Small 4 119B Thinking

Added Mar 17, 2026Open weights

Compare models

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

Loading models…

Use in the app

Mistral Small 4 with reasoning enabled (reasoning_effort=high). A hybrid MoE model with deep step-by-step reasoning for complex prompts, coding, and multi-step problem solving.

  • Reasoning
  • Tools
  • Vision
  • Structured output
Input
Text, Images
Output
Text
Reasoning effort
none · high
Discovery tier
Open weights

Pricing

USD per 1M tokens
Input
$0.40
Output
$1.40

Reference prices as published by the model provider. API example

Specifications

Context window
262.1K
Max output
16.4K
Average output (7 days)
997 tokens
Longer than 70% of 465 models

Side by side

Mistral Small 4 119B Thinking 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 Small 4 119B ThinkingMistralThis modelMistral Small 3.2 24B (2506)MistralRemove
ProviderMistralMistral
Discovery tierOpen weightsOpen weights
AddedMar 17, 2026Apr 15, 2025
Limits
Context window262.14KBest128K
Max output16.38K16.38K
Average output997 tokens—
Pricing, USD per 1M tokens
Input$0.40$0.20Best
Output$1.40$0.40Best
Performance
Throughput123.6 tok/s—
Time to first token1.7 s—
Request success (24h)95.9%100.0%Best
Benchmark indexes
Intelligence Index11.3—
Coding Index26.6—
Capabilities
ReasoningYesNo
ToolsYesNo
VisionYesNo
Structured outputYesNo
Reasoning effortnone, high—

Public evaluations

Benchmarks

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

  • Mistral Small 4 119B Thinking
  • Mistral Small 3.2 24B (2506)

Intelligence Index

  • Mistral Small 4 119B Thinking11.3
  • Mistral Small 3.2 24B (2506)—

Mistral Small 4 119B Thinking: Better than 25% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • Mistral Small 4 119B Thinking: 26.6
    • Mistral Small 3.2 24B (2506): —

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • Mistral Small 4 119B Thinking: 1.0%
    • Mistral Small 3.2 24B (2506): —

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • Mistral Small 4 119B Thinking: 0.0%
    • Mistral Small 3.2 24B (2506): —

  • AA-Briefcasev1.1

    Agentic knowledge work, rated head to head

    • Mistral Small 4 119B Thinking: 241 Elo
    • Mistral Small 3.2 24B (2506): —

  • GDPval-AAv2.1

    Economically valuable occupational tasks, rated head to head

    • Mistral Small 4 119B Thinking: 362 Elo
    • Mistral Small 3.2 24B (2506): —

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • Mistral Small 4 119B Thinking: 0.0%
    • Mistral Small 3.2 24B (2506): —

  • AA-LCRv1.1

    Reasoning across long documents

    • Mistral Small 4 119B Thinking: 49.7%
    • Mistral Small 3.2 24B (2506): —

  • MLCR-AA

    Long-context reasoning over medical records

    • Mistral Small 4 119B Thinking: 1.7%
    • Mistral Small 3.2 24B (2506): —

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • Mistral Small 4 119B Thinking: 9.9%
    • Mistral Small 3.2 24B (2506): —

  • GPQA Diamond

    Graduate-level science questions

    • Mistral Small 4 119B Thinking: 76.9%
    • Mistral Small 3.2 24B (2506): —

  • IFBench

    Following precise output instructions

    • Mistral Small 4 119B Thinking: 48.2%
    • Mistral Small 3.2 24B (2506): —

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • Mistral Small 4 119B Thinking: 0.0%
    • Mistral Small 3.2 24B (2506): —

  • SciCode

    Scientific computing in Python

    • Mistral Small 4 119B Thinking: 38.8%
    • Mistral Small 3.2 24B (2506): —

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • Mistral Small 4 119B Thinking: 17.4%
    • Mistral Small 3.2 24B (2506): —

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • Mistral Small 4 119B Thinking: 41.2%
    • Mistral Small 3.2 24B (2506): —

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • Mistral Small 4 119B Thinking: 49.7%
    • Mistral Small 3.2 24B (2506): —

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.

Throughput
123.6 tok/s
Faster than 70% of 251 models
Time to first token
1.7 s
Quicker than 83% of 251 models
Prompt cache hits
23.6%
Request success (24h)
95.9%
Reliable

Cache hits are the share of input tokens recently served from a prompt cache across all requests. They are not a forecast of your own hit rate.

Hourly request success · last 72 hoursLowest hour 66.6%
  • 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-small-4-119b-2603:thinking",
    "reasoning_effort": "none",
    "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-small-4-119b-2603:thinking",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

Mistral

Other Mistral models, starting with the same family.