Model comparison

DeepSeek-V3.2-Exp vs Devstral Small 2505

DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 34.3 on the Noometry Index. Devstral Small 2505 costs 1.9× less per token, which makes it the better buy when DeepSeek-V3.2-Exp's lead doesn't matter for your workload.

Last verified October 10, 2026. 4 shared benchmarks.

DeepSeek-V3.2-Exp

DeepSeek

44.3

Rank #78 Confirmed

Devstral Small 2505

Mistral AI

34.3

Rank #233 Reported

Summary

They share 4 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 2 categories and Devstral Small 2505 in 0 categories; 2 gaps are clear of the uncertainty.
The widest gap is in coding, where DeepSeek-V3.2-Exp leads 46.5 to 38.9.
The biggest single-benchmark swing is Kagi LLM Benchmark: 52.2% for DeepSeek-V3.2-Exp and 37.7% for Devstral Small 2505.
Devstral Small 2505 is cheaper at $0.10 / $0.30 per million input/output tokens, against $0.26 / $0.38 for DeepSeek-V3.2-Exp.
DeepSeek-V3.2-Exp accepts more context: 164K tokens versus 128K.

Side by side

DeepSeek-V3.2-Exp and Devstral Small 2505 specifications
	DeepSeek-V3.2-Exp	Devstral Small 2505
Provider	DeepSeek	Mistral AI
Noometry Index	44.3	34.3
Released	2025-09-29	2025-05-07
Weights	Open	Open
Context window	164K	128K
Max output	66K	128K
Input $ / M tokens	$0.26	$0.10
Output $ / M tokens	$0.38	$0.30
Results tracked	49	4

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 46.5 (#65), Devstral Small 2505: 38.9 (#166)

Coding benchmarks
Benchmark	DeepSeek-V3.2-Exp	Devstral Small 2505
SWE-bench Verified (bash only)	70%	56.4%
SciCode	38.9%	28.8%
Aider Polyglot	74.2%	—
LMArena WebDev	1362	—
SWE-bench Multilingual	59%	—
WeirdML	39.5%	—
LMArena Coding	1454	—

Agentic & Tool Use Not comparable

DeepSeek-V3.2-Exp: 32.7 (#59), Devstral Small 2505: —

Agentic & Tool Use benchmarks
Benchmark	DeepSeek-V3.2-Exp	Devstral Small 2505
Terminal-Bench	39.6%	—
APEX-Agents	21.3%	—
Berkeley Function Calling Leaderboard	56.7%	—
TheAgentCompany	42.9%	—
Vending-Bench 2	1,034	—

Reasoning DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 22.1 (#208), Devstral Small 2505: 19.7 (#252)

Reasoning benchmarks
Benchmark	DeepSeek-V3.2-Exp	Devstral Small 2505
Kagi LLM Benchmark	52.2%	37.7%
CritPt	2.9%	0%
ARC-AGI-2	4%	—
NYT Connections (extended)	36.7%	—
ARC-AGI-1	57%	—
Chess Puzzles	14%	—
Thematic Generalization	65%	—
LMArena Hard Prompts	1434	—
DTBench	87.7%	—
LMCA	29.1%	—
Epoch Capabilities Index	146.27	—

Math Not comparable

DeepSeek-V3.2-Exp: 41.7 (#87), Devstral Small 2505: —

Math benchmarks
Benchmark	DeepSeek-V3.2-Exp	Devstral Small 2505
MathArena Final-Answer Competitions	57.7%	—
OTIS Mock AIME 2024-2025	87.8%	—
ProofBench	8%	—
LMArena Math	1435	—
FrontierMath (Feb 2025 set)	22.1%	—
FrontierMath Tier 4 (v1)	2.1%	—

Knowledge Not comparable

DeepSeek-V3.2-Exp: 51.7 (#66), Devstral Small 2505: —

Knowledge benchmarks
Benchmark	DeepSeek-V3.2-Exp	Devstral Small 2505
GPQA Diamond	83.4%	—
Vectara Hallucination Rate	5.3%	—
LMArena Expert	1436	—

Multilingual Not comparable

DeepSeek-V3.2-Exp: 52.2 (#90), Devstral Small 2505: —

Multilingual benchmarks
Benchmark	DeepSeek-V3.2-Exp	Devstral Small 2505
LMArena Non-English	1409	—
LMArena Chinese	1461	—
LMArena French	1433	—
LMArena German	1440	—
LMArena Japanese	1374	—
LMArena Korean	1371	—
LMArena Russian	1424	—
LMArena Spanish	1440	—

Instruction Following Not comparable

DeepSeek-V3.2-Exp: 74.5 (#93), Devstral Small 2505: —

Instruction Following benchmarks
Benchmark	DeepSeek-V3.2-Exp	Devstral Small 2505
LMArena Instruction Following	1413	—

Long Context Not comparable

DeepSeek-V3.2-Exp: 47.6 (#16), Devstral Small 2505: —

Long Context benchmarks
Benchmark	DeepSeek-V3.2-Exp	Devstral Small 2505
Fiction.LiveBench	83.3%	—
CL-bench	13.2%	—
CL-bench Life	9.5%	—
LMArena Longer Query	1428	—

Writing & Preference Not comparable

DeepSeek-V3.2-Exp: 62.4 (#77), Devstral Small 2505: —

Writing & Preference benchmarks
Benchmark	DeepSeek-V3.2-Exp	Devstral Small 2505
LMArena Text	1425	—
LMArena Creative Writing	1403	—
EQ-Bench Creative Writing	1515	—
LMArena Multi-Turn	1427	—

Frequently asked questions

Is DeepSeek-V3.2-Exp better than Devstral Small 2505?

Which is cheaper, DeepSeek-V3.2-Exp or Devstral Small 2505?

Devstral Small 2505 is cheaper. It lists at $0.10 per million input tokens and $0.30 per million output tokens; DeepSeek-V3.2-Exp lists at $0.26 and $0.38.

Is DeepSeek-V3.2-Exp or Devstral Small 2505 better for coding?

DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 versus 38.9 in the Noometry coding category.

Which has the bigger context window?

DeepSeek-V3.2-Exp does, with 164K tokens against 128K.

How many benchmarks do DeepSeek-V3.2-Exp and Devstral Small 2505 share?

4 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 49 scored results on Noometry and Devstral Small 2505 has 4.