Devstral Small 2

Devstral Small 2

Mistral · Released Dec 9, 2025
Intelligence #402 / 620
25.9 our score
Speed #185 / 289
64.8 tok/s
Input Price
– Not priced
Output Price
– Not priced
Context
– Not reported

Analysis Summary

Devstral Small 2 is Mistral's coding-oriented small model, released December 2025. Its intelligence index of 17.4 and coding index of 29.3 place it above the general Ministral 3 family on technical tasks. The terminal bench score of 0.167 and livecodebench of 0.348 confirm a meaningful coding focus relative to its size. Agentic performance at 20.0 is limited but functional for simple tool-use scenarios.

For businesses, this model is best suited to lightweight coding assistance, code review, and simple software engineering tasks where cost is a constraint. It is not a replacement for frontier coding models in complex agentic pipelines, but it can handle routine code generation and debugging tasks reliably. Instruction-following and long-context scores are modest, so complex multi-file tasks may be unreliable.

No pricing data is available for this variant. Teams evaluating small coding models should compare it against similarly sized open-weight alternatives. For organisations already using Mistral's API, it is a practical low-cost option for coding-adjacent workloads that do not require frontier-level capability.

Assessed July 10, 2026

Editorial notes

Devstral Small 2 from Mistral is a coding-focused small model with a 29-point coding index and strong terminal bench performance, suited to lightweight software engineering tasks at low cost.

Rankings consider pricing, capabilities, benchmarks, and real-world applicability and are refreshed as new models launch. Feedback?

DFO Verdict

Devstral Small 2 from Mistral is a coding-focused small model with a 29-point coding index and strong terminal bench performance, suited to lightweight software engineering tasks at low cost.

#402 of 620 overall

Benchmark scores

GPQA Diamond 53.2%
HLE 3.4%
MMLU Pro 67.8%
AIME 2025 34.3%
SciCode 28.8%
LiveCodeBench 34.8%
TerminalBench Hard 16.7%
τ²-Bench 23.4%
IFBench 31.2%
LCR 24%

Magenta = intelligence Ā· Ink = technical/agentic Ā· Cyan = content & long-context Ā· Grey = community benchmarks. Data: Artificial Analysis, Hugging Face.

17.4 Intelligence IndexĀ·29.3 Coding IndexĀ·20 Agentic IndexĀ·34.3 Math Index

How Devstral Small 2 compares

Devstral Small 2 ranks #169 of 395 AI models we track for overall intelligence, #79 of 160 for coding, #176 of 302 for agentic tasks. Devstral Small 2 is currently free to use via OpenRouter.

Position in the field
Intelligence: smarter than 35% of models #402
Speed: faster than 36% of models #185
worst in fieldmedianbest in field
Price vs frontier peers Ā· $ per 1M tokens
Devstral Small 2 $0.00 in $0.00 out
Anthropic: Claude Fable 5 $10.00 in $50.00 out
Anthropic: Claude Opus 4.8 $5.00 in $25.00 out
Claude Opus 5 $5.00 in $25.00 out

Dark bar = input Ā· light bar = output, scaled to the priciest peer.

Context window vs peers Ā· tokens

1M tokens ā‰ˆ 8 full-length novels or ~2,500 pages of business documents in a single request.

Intelligence3Technical3Value0Content5
Performance profile

Strongest on content. The pulled-in value corner is the trade-off, and if the shape matters more than the price, this is your model.

Compare shapes side-by-side →

Frequently asked questions about Devstral Small 2

How much does Devstral Small 2 cost?

Devstral Small 2 is currently available for free via OpenRouter.

Is Devstral Small 2 good for coding?

On our coding benchmark index, Devstral Small 2 ranks #79 of 160 models, placing it in the broader range of the field for code generation and debugging.

Who created Devstral Small 2?

Devstral Small 2 is developed by Mistral and was released on December 9, 2025.