lovedheart / deepseek-v3 family
deepseek-v3.2-exp
deepseek/deepseek-v3.2-exp
Benchmarks
Where it scores.
General
82.6
Coding
66.9
Math
69.8
Reasoning
60.2
Normalized 0–100 across indexed models.
Runnable variants
Pick a size.
No recommended variants are indexed for this model.
Raw benchmarks
The inputs.
| Source | Task | Raw | Normalized |
|---|---|---|---|
| Arena | ArenaScore | 1421.7 | 90.2 |
| ZeroEval | AIME2025 | 89.3 | 89.3 |
| ZeroEval | GPQA | 79.9 | 79.9 |
| ZeroEval | HLE | 19.8 | 19.8 |
| ZeroEval | SWEBenchVerified | 67.8 | 67.8 |
| ZeroEval | SimpleQA | 97.1 | 97.1 |
From the Hugging Face model card
About this model.
From the Hugging Face model card