Neura Intelligence Index
Back to Leaderboard
DeepSeek-V3.2-Speciale
🇨🇳DeepSeek7moOpen Weights
Parameters:685000000000B
Neura Index
60.0
Rank #109—
mediumCategory Profile
Benchmark Scores (4)
math
| Benchmark | Score | Source |
|---|---|---|
| AIME 2025 | 96 | verified |
reasoning
| Benchmark | Score | Source |
|---|---|---|
| Humanity's Last Exam | 30.6 | verified |
coding
| Benchmark | Score | Source |
|---|---|---|
| SWE-bench Verified | 73.1 | verified |
agent
| Benchmark | Score | Source |
|---|---|---|
| Toolathlon | 35.2 | verified |
Performance History
Similar Models
Frequently Asked Questions about DeepSeek-V3.2-Speciale
What is DeepSeek-V3.2-Speciale's Neura Intelligence Index score?▾
DeepSeek-V3.2-Speciale by DeepSeek scores 60.0 out of 100 on the Neura Intelligence Index, ranking #109 among tracked AI models. This composite score is computed from 4 individual benchmarks across coding, math, reasoning, and language categories.
Is DeepSeek-V3.2-Speciale open source?▾
Yes, DeepSeek-V3.2-Speciale has open weights available under the mit license. You can download and self-host the model.
What benchmark does DeepSeek-V3.2-Speciale perform best on?▾
DeepSeek-V3.2-Speciale achieves its highest score on AIME 2025 with a score of 96.0%. This benchmark evaluates american invitational mathematics examination 2025 — competition-level math problems.