DeepSeek V3.2 Speciale
DeepSeekDeepSeekOpen WeightMIT · Commercial OK
Description
DeepSeek-V3.2-Speciale is a specialized variant of DeepSeek-V3.2, optimized for enhanced performance on specific tasks.
Release Date
2025-12-01
Parameters
685.0B
Context Length
164K
Modalities
text
Capability Radar
46
general
57
coding
95
reasoning
60
scienceest.
50
agents
0
multimodal
Science uses a reasoning proxy when dedicated science benchmarks are unavailable.
Rankings
| Domain | #Rank | Score | Source |
|---|---|---|---|
| Agents & Tools | 66 | 53.0 | LS |
| Code Ranking | 51 | 72.0 | AA |
| General Ranking | 185 | 52.0 | AA |
| Math Reasoning | 5 | 98.0 | AA |
| Science | 37 | 75.0 | AA |
Benchmark Scores (LLM Stats)
Agents
t2-bench
80.3%SR
Terminal-Bench 2.0
46.4%SR
Toolathlon
35.2%SR
Code
SWE-Bench Verified
73.1%SR
Math
HMMT 2025
99.2%SR
AIME 2025
96.0%SR
CodeForces
0.90 / 3000SR
Humanity's Last Exam
30.6%SR
AA Evaluation Indices
Math Index96.7
Coding Index37.9
Intelligence Index29.4
Aime 251.0
Livecodebench0.9
Gpqa0.9
Mmlu Pro0.9
Ifbench0.6
Lcr0.6
Scicode0.4
Terminalbench Hard0.3
Hle0.3
Tau20.0
LLM Stats Category Scores
Math80
Frontend Development70
Code60
Reasoning60
Tool Calling50
Agents50
Vision30
Pricing
Input PriceFree
Output PriceFree
Blended Price (3:1)Free
Speed
Tokens/sec0.0 tokens/s
Time to First Token0.00s
Time to Answer0.00s
Available Providers
(LS internal units)No provider data available