Inference
EN
Mistral

Devstral Small

Devstral Small is a standard model from Mistral. 8 benchmarks count toward its index, in 5 categories.

  • Open weights
  • Standard
  • Input: text
  • Released 2025-05
RankedEnough results for an index.
IndexAll categories, out of 100.28.2±8.2Median 51.0
CoverageIndex weight that has results.75%
SpeedOutput tokens per second.—Median 67/s
First tokenSeconds to the first output token.—Median 1.3 s
Input / 1MUS dollars per 1M input tokens.—Median $0.60
Output / 1MUS dollars per 1M output tokens.—Median $2.55
ContextMaximum tokens in one request.128KMedian 262K
EloLMArena rating and rank.—Median 1422

Run locally

24.01B parameters
16-bit
51.4 GB
8-bit
27.8 GB
4-bit
16.3 GB
Weights
Hugging Face (opens in a new tab) ↗

Memory need with a short context. The 4-bit copy fits in 24 GB. Compare local models

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.1 counted21.6—
CodingCode writing and repair.1 counted35.7—
ReasoningLogic problems and puzzles.2 counted28.5211 / 227
MultimodalTasks with images and text.None counted——
KnowledgeFacts and expert knowledge.3 counted24.2260 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted21.4—
MathMath problems.None counted——

Results

8 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
Critical Physics TasksReasoning0.0%150 / 214
τ²-Bench Tool-Agent-User EvaluationAgentic28.4%115 / 149
Artificial Analysis IFBenchInstruction34.6%123 / 146
Artificial Analysis Long Context ReasoningReasoning18.7%194 / 217
Software Engineering Benchmark VerifiedCoding46.8%103 / 114
Artificial Analysis Humanity's Last ExamKnowledge3.8%202 / 222
Artificial Analysis Omniscience AccuracyKnowledge14.7%153 / 164
Artificial Analysis GPQA DiamondKnowledge41.4%189 / 199

Sources

BenchLMCC BY-NC 4.0

History

5 releases
Index28.8 → 28.2
23 Oct 202612 Nov 2026

Questions

How good is Devstral Small?
Its index is 28.2 out of 100. That is place 255 of 283 ranked models.
What is Devstral Small best at?
Reasoning: 28.5 out of 100, place 211 of 227.
What is the context window of Devstral Small?
128K tokens.
Is Devstral Small open source?
Yes. Its weights are published.
When was Devstral Small released?
May 2025.

Badge

Devstral Small on Inference360: place 255 of 283, index 28.2
HTML
<a href="https://inference360.ai/models/devstral-small"><img src="https://inference360.ai/badges/devstral-small.svg" alt="Devstral Small on Inference360: place 255 of 283, index 28.2" height="20"></a>
Markdown
[![Devstral Small on Inference360: place 255 of 283, index 28.2](https://inference360.ai/badges/devstral-small.svg)](https://inference360.ai/models/devstral-small)

Compare with

Gemma 4 E4B28.2Llama 3.3 70B28.2GPT-4o28.7Mistral Medium 327.1Magistral Medium33.7

More from Mistral

All models
Mistral Medium 327.1Magistral Medium33.7Magistral Small31.4Mistral Small 3.2 24BUnrankedMistral Small 3.1 24BUnranked