Deepseek V3.2 Speciale
DeepSeek develops the DeepSeek V 3.2 Speciale, a chat model exceling at code generation, function calling, and reasoning, with notable strengths in coding and mathematical tasks. Its technical architecture features DeepSeek Sparse Attention, an efficient attention mechanism optimized for long-context scenarios, supporting a context window of 163,840 tokens.
Input
Output
Context
164K
Max Output
64K
Parameters
685.4B
Technical Specifications
Model TypeChat
Context Window163,840 tokens
Max Output Tokens64,000 tokens
Parameters685.4B
Release DateNov 28, 2025
Training CutoffNot available
Licensemit
Open SourceYes
Input Modalities
Text
Output Modalities
Text
Capabilities
Benchmarks
Artificial Analysis
26.1%
HLE
87.1%
GPQA
44.0%
SciCode
86.3%
MMLU Pro
96.7
Math
37.9
Coding
89.6%
LiveCodeBench
22.2
Intelligence
Resources & Links
Related Tools
Data sourced from official provider APIs and documentation
Last updated: Jun 23, 2026
Start building with the right model.
Automatically route workloads to the right model for every task, every time.