modelgrep
M

Mistral: Mistral Small 4

mistralai/mistral-small-2603

127th smartest of 178ReasoningToolsJSONVision
Use via OpenRouter ↗
Intelligence
18.6
127th of 178
Design Elo
Speed
117
52nd fastest
Latency
369ms
first token
Input price
$0.150
78th cheapest
Context
262K

How it compares

Smarter than29%
of all ranked models
Faster than82%
of all ranked models
Cheaper than74%
of all ranked models

Overview

Mistral Small 4 is the next major release in the Mistral Small family, unifying the capabilities of several flagship Mistral models into a single system. It combines strong reasoning from...

Benchmarks

independent · via OpenRouter
Artificial Analysis28th percentile
Intelligence Index
18.6
Coding Index
16.4
Agentic Index
18.6
GPQA Diamond
57%
Humanity's Last Exam
4%
SciCode
28%
Tau²-Bench (agentic)
18%

Providers & pricing (2)

ProviderIn $/MOut $/MUptime
Mistral$0.150$0.600100%
Venicefp8$0.188$0.75096.5%

Specifications

Context window262K
Max output
Knowledge cutoff
Input modalitiestext, image
Output modalitiestext
Prompt caching
Cache read price$0.015/M
ModeratedNo

Mistral Small 4 FAQ

How much does Mistral Small 4 cost?

Mistral Small 4 costs $0.150 per million input tokens and $0.600 per million output tokens via OpenRouter, making it 78th cheapest of 298 paid models.

How smart is Mistral Small 4?

Mistral Small 4 scores 18.6 on the Artificial Analysis Intelligence Index, ranking 127th of 178 benchmarked models, with a GPQA Diamond score of 57%.

How fast is Mistral Small 4?

Mistral Small 4 generates around 117 tokens per second with 369ms time-to-first-token (p50), the 52nd fastest tracked model.

What is Mistral Small 4's context window?

Mistral Small 4 supports a 262K-token context window. It accepts text, image input.

Compare head-to-head