The AI arena is free today

Open Superagent

DeepSeek-V3.2-Speciale vs Mistral Large 3 (675B Instruct 2512 NVFP4)

DeepSeek-V3.2-Speciale leads the LLM Stats Score 33.9 to 8.9.

DeepSeek · Mistral AI · Updated for 2026

Which is better?

DeepSeek-V3.2-Speciale leads the overall LLM Stats Score 33.9 to 8.9, ranking #99 overall.

Based on current LLM Stats indexes, shared benchmarks, pricing, and model metadata for 2026.

Choose DeepSeek-V3.2-Speciale

  • overall performance matters — it scores 33.9 and ranks #99 on LLM Stats
  • your work emphasizes reasoning and coding — it leads those capability indexes

Choose Mistral Large 3 (675B Instruct 2512 NVFP4)

  • you want the most recent training data — it shipped Dec 2025

At a glance

The differences that matter most.

DeepSeek-V3.2-Speciale
DeepSeek
Mistral Large 3 (675B Instruct 2512 NVFP4)
Mistral AI
Core performance indexes
LLM Stats Score
33.9
#99
8.9
#277
Reasoning
32.5
#106
9.1
#270
Coding
18.6
#114
1.4
#235
Cost, coverage & limits
Benchmark wins
—
—
Input price
$0.28 / M
— / M
Output price
$0.42 / M
— / M
Context window
131,072
—

Capability indexes

Additional strengths measured across groups of related public benchmarks

1 shared
Index
DeepSeek-V3.2-Speciale
Mistral Large 3 (675B Instruct 2512 NVFP4)
Math
34.2#48
18.9#175
Conservative TrueSkill rating · higher is betterHow scores work

Individual benchmarks

8 reported for DeepSeek-V3.2-Speciale · 5 for Mistral Large 3 (675B Instruct 2512 NVFP4)

No common benchmarks found

DeepSeek-V3.2-Speciale and Mistral Large 3 (675B Instruct 2512 NVFP4)don't have any common benchmark datasets to compare. They may have been evaluated on different testing suites.

Human preference

Blind head-to-head votes and playground preference scores

Model Size

Parameter count comparison

10.0B diff

DeepSeek-V3.2-Speciale has 10.0B more parameters than Mistral Large 3 (675B Instruct 2512 NVFP4), making it 1.5% larger.

DeepSeek
DeepSeek-V3.2-Speciale
685.0Bparameters
Mistral AI
Mistral Large 3 (675B Instruct 2512 NVFP4)
675.0Bparameters
685.0B
DeepSeek-V3.2-Speciale
675.0B
Mistral Large 3 (675B Instruct 2512 NVFP4)

Context Window

Maximum input and output token capacity

Only DeepSeek-V3.2-Speciale specifies input context (131,072 tokens). Only DeepSeek-V3.2-Speciale specifies output context (131,072 tokens).

DeepSeek
DeepSeek-V3.2-Speciale
Input131,072 tokens
Output131,072 tokens
Mistral AI
Mistral Large 3 (675B Instruct 2512 NVFP4)
Input- tokens
Output- tokens
Tue Sep 22 2026 • llm-stats.com

Input capabilities

Documented input modalities across available providers

Mistral Large 3 (675B Instruct 2512 NVFP4) supports multimodal inputs, whereas DeepSeek-V3.2-Speciale does not.

Mistral Large 3 (675B Instruct 2512 NVFP4) can handle both text and other forms of data like images, making it suitable for multimodal applications.

DeepSeek-V3.2-Speciale

Text
Images
Audio
Video

Mistral Large 3 (675B Instruct 2512 NVFP4)

Text
Images
Audio
Video

License

Usage and distribution terms

DeepSeek-V3.2-Speciale is licensed under MIT, while Mistral Large 3 (675B Instruct 2512 NVFP4) uses Apache 2.0.

License differences may affect how you can use these models in commercial or open-source projects.

DeepSeek-V3.2-Speciale

MIT

Open weights

Mistral Large 3 (675B Instruct 2512 NVFP4)

Apache 2.0

Open weights

Release Timeline

When each model was launched

DeepSeek-V3.2-Speciale was released on 2025-12-01, while Mistral Large 3 (675B Instruct 2512 NVFP4) was released on 2025-12-04.

Mistral Large 3 (675B Instruct 2512 NVFP4) is 0 month newer than DeepSeek-V3.2-Speciale.

DeepSeek-V3.2-Speciale

Dec 1, 2025

9 months ago

Mistral Large 3 (675B Instruct 2512 NVFP4)

Dec 4, 2025

9 months ago

3d newer

Knowledge Cutoff

When training data ends

Neither model specifies a knowledge cutoff date.

Unable to compare the recency of their training data.

No cutoff dates available

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion→

Judge for yourself.

Run your own prompts against DeepSeek-V3.2-Speciale and Mistral Large 3 (675B Instruct 2512 NVFP4) side-by-side, then vote on the output you prefer.

DeepSeek-V3.2-Speciale
✓ Preferred
Mistral Large 3 (675B Instruct 2512 NVFP4)
Open in Playground

FAQ

Common questions about DeepSeek-V3.2-Speciale vs Mistral Large 3 (675B Instruct 2512 NVFP4).

Which is better, DeepSeek-V3.2-Speciale or Mistral Large 3 (675B Instruct 2512 NVFP4)?

DeepSeek-V3.2-Speciale leads the LLM Stats Score 33.9 to 8.9. DeepSeek-V3.2-Speciale is made by DeepSeek and Mistral Large 3 (675B Instruct 2512 NVFP4) is made by Mistral AI. The best choice depends on your use case — compare their capability indexes, individual benchmarks, pricing, and limits above.

How does DeepSeek-V3.2-Speciale compare to Mistral Large 3 (675B Instruct 2512 NVFP4) in benchmarks?

DeepSeek-V3.2-Speciale scores HMMT 2025: 99.2%, AIME 2025: 96.0%, CodeForces: 90.0%, t2-bench: 80.3%, SWE-Bench Verified: 73.1%. Mistral Large 3 (675B Instruct 2512 NVFP4) scores MMMLU: 85.5%, AMC_2022_23: 52.0%, GPQA: 43.9%, LiveCodeBench: 34.4%, SimpleQA: 23.8%.

What are the context window sizes for DeepSeek-V3.2-Speciale and Mistral Large 3 (675B Instruct 2512 NVFP4)?

DeepSeek-V3.2-Speciale supports 131K tokens and Mistral Large 3 (675B Instruct 2512 NVFP4) supports an unknown number of tokens. A larger context window lets you process longer documents, conversations, or codebases in a single request.

What are the main differences between DeepSeek-V3.2-Speciale and Mistral Large 3 (675B Instruct 2512 NVFP4)?

Key differences include LLM Stats Score (33.9 vs 8.9), multimodal support (no vs yes), licensing (MIT vs Apache 2.0). See the full comparison above for benchmark-by-benchmark results.

Who makes DeepSeek-V3.2-Speciale and Mistral Large 3 (675B Instruct 2512 NVFP4)?

DeepSeek-V3.2-Speciale is developed by DeepSeek and Mistral Large 3 (675B Instruct 2512 NVFP4) is developed by Mistral AI.

Related comparisons