Model comparison

DeepSeek-V3.2-Speciale vs Ministral 8B

DeepSeek-V3.2-Speciale is the stronger model overall, scoring 39.7 to 28.2 on the Noometry Index. Ministral 8B costs 5.7× less per token, which makes it the better buy when DeepSeek-V3.2-Speciale's lead doesn't matter for your workload.

Last verified October 10, 2026. 0 shared benchmarks.

DeepSeek-V3.2-Speciale

DeepSeek

39.7

Rank #162 Reported

Ministral 8B

Mistral AI

28.2

Rank #325 Confirmed

Summary

The widest gap is in reasoning, where DeepSeek-V3.2-Speciale leads 32.9 to 18.4.
Ministral 8B is cheaper at $0.15 / $0.15 per million input/output tokens, against $0.58 / $1.68 for DeepSeek-V3.2-Speciale.
Ministral 8B accepts more context: 262K tokens versus 128K.

Side by side

DeepSeek-V3.2-Speciale and Ministral 8B specifications
	DeepSeek-V3.2-Speciale	Ministral 8B
Provider	DeepSeek	Mistral AI
Noometry Index	39.7	28.2
Released	2025-12-01	2024-10-01
Weights	Open	Open
Context window	128K	262K
Max output	128K	262K
Input $ / M tokens	$0.58	$0.15
Output $ / M tokens	$1.68	$0.15
Results tracked	3	17

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

Category by category

Coding DeepSeek-V3.2-Speciale leads

DeepSeek-V3.2-Speciale: 40.4 (#140), Ministral 8B: 35.0 (#230)

Coding benchmarks
Benchmark	DeepSeek-V3.2-Speciale	Ministral 8B
WeirdML	46.7%	—
LMArena Coding	—	1202

Agentic & Tool Use Not comparable

DeepSeek-V3.2-Speciale: —, Ministral 8B: 16.4 (#148)

Agentic & Tool Use benchmarks
Benchmark	DeepSeek-V3.2-Speciale	Ministral 8B
Berkeley Function Calling Leaderboard	—	11.1%

Reasoning DeepSeek-V3.2-Speciale leads

DeepSeek-V3.2-Speciale: 32.9 (#73), Ministral 8B: 18.4 (#281)

Reasoning benchmarks
Benchmark	DeepSeek-V3.2-Speciale	Ministral 8B
SimpleBench	52.6%	—
LMArena Hard Prompts	—	1191
DTBench	—	45.7%

Math Not comparable

DeepSeek-V3.2-Speciale: —, Ministral 8B: 25.7 (#267)

Math benchmarks
Benchmark	DeepSeek-V3.2-Speciale	Ministral 8B
LMArena Math	—	1188
MATH Level 5	—	14.9%

Knowledge Not comparable

DeepSeek-V3.2-Speciale: —, Ministral 8B: 12.6 (#297)

Knowledge benchmarks
Benchmark	DeepSeek-V3.2-Speciale	Ministral 8B
GPQA Diamond	—	27.1%
Vectara Hallucination Rate	—	7.4%
LMArena Expert	—	1170

Multilingual Not comparable

DeepSeek-V3.2-Speciale: —, Ministral 8B: 35.1 (#247)

Multilingual benchmarks
Benchmark	DeepSeek-V3.2-Speciale	Ministral 8B
LMArena Non-English	—	1165
LMArena Chinese	—	1193
LMArena Russian	—	1195

Instruction Following Not comparable

DeepSeek-V3.2-Speciale: —, Ministral 8B: 60.5 (#250)

Instruction Following benchmarks
Benchmark	DeepSeek-V3.2-Speciale	Ministral 8B
LMArena Instruction Following	—	1161

Long Context Not comparable

DeepSeek-V3.2-Speciale: —, Ministral 8B: 36.7 (#227)

Long Context benchmarks
Benchmark	DeepSeek-V3.2-Speciale	Ministral 8B
LMArena Longer Query	—	1212

Writing & Preference DeepSeek-V3.2-Speciale leads

DeepSeek-V3.2-Speciale: 46.0 (#222), Ministral 8B: 39.6 (#246)

Writing & Preference benchmarks
Benchmark	DeepSeek-V3.2-Speciale	Ministral 8B
LMArena Text	—	1191
LMArena Creative Writing	—	1175
EQ-Bench Creative Writing	1276	—
LMArena Multi-Turn	—	1166

Frequently asked questions

Is DeepSeek-V3.2-Speciale better than Ministral 8B?

Which is cheaper, DeepSeek-V3.2-Speciale or Ministral 8B?

Ministral 8B is cheaper. It lists at $0.15 per million input tokens and $0.15 per million output tokens; DeepSeek-V3.2-Speciale lists at $0.58 and $1.68.

Is DeepSeek-V3.2-Speciale or Ministral 8B better for coding?

DeepSeek-V3.2-Speciale scores higher on coding benchmarks: 40.4 versus 35.0 in the Noometry coding category.

Which has the bigger context window?

Ministral 8B does, with 262K tokens against 128K.

How many benchmarks do DeepSeek-V3.2-Speciale and Ministral 8B share?

0 benchmarks have published results for both models. DeepSeek-V3.2-Speciale has 3 scored results on Noometry and Ministral 8B has 17.