Inference
EN
Mistral

Mistral Small 4

Mistral Small 4 is a standard model from Mistral. 12 benchmarks count toward its index, in 6 categories.

  • Open weights
  • Standard
  • Input: text · image
  • Released 2026-03
RankedEnough results for an index.
IndexAll categories, out of 100.37.9±6.2Median 51.0
CoverageIndex weight that has results.85%
SpeedOutput tokens per second.114/sMedian 67/s
First tokenSeconds to the first output token.0.6 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$0.15batch $0.075Median $0.60
Output / 1MUS dollars per 1M output tokens.$0.60batch $0.30 Batch price per 1M output tokens.Median $2.55
ContextMaximum tokens in one request.262KMedian 262K
EloLMArena rating and rank.—Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
mistralai/mistral-small-2603
Max outputMaximum output tokens per request.
210K
Knowledge cutoffLast date of training data.
—
ToolsTool calls on OpenRouter.
Yes
JSON outputOutput follows a given JSON schema.
Yes
Cache read / 1MUS dollars per 1M cached tokens.
$0.015
WeightsLink to the published weights.
Hugging Face (opens in a new tab) ↗
RetiresDate OpenRouter removes the model.
—

Where to run

APIs that serve this model.
3 providers
ProviderTypeMaker, cloud, host or router.InputOutputContext
OpenRouterRouter$0.15$0.60262K
SAP AI CoreDeprecatedCloud$0.15$0.60128K
RegoloHost$0.561$2.36256K

US dollars per 1M tokens, list price. From models.dev (opens in a new tab) (MIT).

Run locally

119.4B parameters
16-bit
251.7 GB
8-bit
134.2 GB
4-bit
77.0 GB
Weights
Hugging Face (opens in a new tab) ↗

Memory need with a short context. The 4-bit copy fits in 128 GB. Compare local models

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.3 counted33.6192 / 218
CodingCode writing and repair.2 counted44.3166 / 209
ReasoningLogic problems and puzzles.2 counted39.4167 / 227
MultimodalTasks with images and text.1 counted29.1—
KnowledgeFacts and expert knowledge.3 counted39.9194 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted34.1—
MathMath problems.None counted——

Results

12 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
Artificial Analysis GPQA DiamondKnowledge76.9%109 / 199
Artificial Analysis IFBenchInstruction48.2%83 / 146
Critical Physics TasksReasoning0.3%137 / 214
Artificial Analysis Humanity's Last ExamKnowledge9.9%145 / 222
τ²-Bench Tool-Agent-User EvaluationAgentic41.2%104 / 149
Artificial Analysis Long Context ReasoningReasoning49.7%157 / 217
Artificial Analysis Omniscience AccuracyKnowledge21.4%124 / 164
GDPval-AA normalizedAgentic0.0%101 / 133
Artificial Analysis SciCodeCoding38.8%93 / 114
Artificial Analysis MMMU-ProMultimodal56.8%94 / 113
Terminal-Bench 2.0 · Vals AIAgentic16.9% ±4.061 / 66
Vibe Code Bench v1.1Coding0.0%120 / 123
SAGEEducation59.2% ±2.129 / 29

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseVals AINo licence statedmodels.devMIT

History

5 releases
Index38.7 → 37.9
23 Oct 202612 Nov 2026
Output / 1M$0.60 → $0.60
23 Oct 202612 Nov 2026

Questions

How much does Mistral Small 4 cost?
$0.15 per 1M input tokens, $0.60 per 1M output tokens.
How good is Mistral Small 4?
Its index is 37.9 out of 100. That is place 222 of 283 ranked models.
What is Mistral Small 4 best at?
Knowledge: 39.9 out of 100, place 194 of 268.
How fast is Mistral Small 4?
114 output tokens per second, 0.6 s to the first token.
What is the context window of Mistral Small 4?
262K tokens.
Is Mistral Small 4 open source?
Yes. Its weights are published.
When was Mistral Small 4 released?
March 2026.

Badge

Mistral Small 4 on Inference360: place 222 of 283, index 37.9
HTML
<a href="https://inference360.ai/models/mistral-small-4"><img src="https://inference360.ai/badges/mistral-small-4.svg" alt="Mistral Small 4 on Inference360: place 222 of 283, index 37.9" height="20"></a>
Markdown
[![Mistral Small 4 on Inference360: place 222 of 283, index 37.9](https://inference360.ai/badges/mistral-small-4.svg)](https://inference360.ai/models/mistral-small-4)

Compare with

DeepSeek V3.137.9North Mini Code38.0Claude 3.7 Sonnet37.7Granite 4.2 8B37.6Mistral Large 334.3

Lower price

dots3-note Preview63.4 · FreeApodex 1.1 Mini55.7 · FreeNemotron 3 Ultra50.8 · Free

More from Mistral

All models
LeanstralUnrankedMistral Small 4 (Reasoning)UnrankedVoxtral Mini 4B Realtime 2602UnrankedMistral Medium 3.5 128B50.3Mistral Medium 3.542.7