Devstral Small 2

Mistral·Released 2025-12-09

Open Source

Compare

Quality Index

19.5

193rd of 440

Top 44%

Coding Index

20.7

142nd of 350

Top 41%

Math Index

34.3

175th of 268

Top 66%

Price/1M

$0.00

3rd cheapest

100% below median

Top 27%

Speed

207 tok/s

Top 7%

TTFT

0.35s

Market Position

Devstral Small 2Market Average

Pricing

Input

$0.00

per 1M tokens

Output

$0.00

per 1M tokens

Blended

$0.00

per 1M tokens

Cheaper than 73% of models. Median price is $0.31/1M tokens.

Cost Calculator

Tokens per day1M

100K100M

Daily

$0.00

Monthly

$0.00

vs. Similar Models

Gemini 2.0 Flash Thinking Experimental (Jan '25)Q:+0.1

$0.00

Gemini 2.5 Flash-Lite Preview (Sep '25) (Non-reasoning)Q:-0.1

$0.17

QwQ 32BQ:+0.2

$0.74

Qwen3 VL 30B A3B (Reasoning)Q:+0.2

$0.75

Performance

207

tokens/sec

Faster than 93% of models

0.35

seconds

Faster than 56% of models

Market Median

46 tok/s

351% faster

Median TTFT

0.43s

18% faster

Speed Comparison

GPT-5.4 nano (Non-Reasoning)

205 tok/s-1%

GPT-5.4 nano (xhigh)

211 tok/s+2%

Nova 2.0 Omni (Non-reasoning)

212 tok/s+2%

Benchmarks

MMLU-Pro

67.8%

GPQA Diamond

53.2%

HLE

3.4%

LiveCodeBench

34.8%

SciCode

28.8%

TerminalBench Hard

16.7%

MATH-500Not evaluated

AIMENot evaluated

AIME 2025

34.3%

IFBench

31.2%

Long Context Recall

24.0%

Tau2

23.4%

Market AverageTop Score

Open Source

Quick Compare

Similar Models

Gemini 2.5 Flash-Lite Preview (Sep '25) (Non-reasoning)

Google

Q: 19.4$0.17/1M1.0M ctx

Faster: 36%Coding: -6.2

Gemini 2.0 Flash Thinking Experimental (Jan '25)

Google

Q: 19.6$0.00/1M

Slower: 100%Coding: +3.4

QwQ 32B

Alibaba

Q: 19.7$0.74/1M131K ctx

Slower: 84%

Qwen3 VL 30B A3B (Reasoning)

Alibaba

Q: 19.7$0.75/1M

Slower: 38%Coding: -7.6

Qwen3 235B A22B (Reasoning)

Alibaba

Q: 19.8$2.63/1M

Slower: 71%Coding: -3.3

Motif-2-12.7B-Reasoning

Motif Technologies

Q: 19.1$0.00/1M

Slower: 100%Coding: -8.8

Compare all 7 models