DeepSeek-V3.2-Exp is an experimental large language model released by DeepSeek as an intermediate step between V3.1 and future architectures. It introduces DeepSeek Sparse Attention (DSA), a fine-grained sparse attention mechanism...
信号概览
评分详解
| 信号 | 强度 | 权重 | 影响 |
|---|---|---|---|
| Benchmarksjust now | 70 | 30% | +20.9 |
| Pricingjust now | 100 | 15% | +14.9 |
| Capabilitiesjust now | 67 | 20% | +13.3 |
| Recencyjust now | 68 | 15% | +10.1 |
| Context Windowjust now | 83 | 10% | +8.3 |
| Output Capacityjust now | 77 | 10% | +7.7 |
Benchmark Performance
LMSYS Arena Elo
1422
Percentile
87.0
Weight
30%
No task benchmark data available yet for this model.
功能
模态
DeepSeek近期发布时间线
把当前模型放回同一服务商最近的发布节奏中查看。
DeepSeek V4.1 Flash
coding
DeepSeek V4.1 Flash (batch)
coding
DeepSeek V4 Flash Vision Exp
coding
DeepSeek V4 Pro 0813
coding
DeepSeek V4 Flash 0731
coding
DeepSeek V4 Pro 0423
coding
DeepSeek V4 Flash 0423
coding
DeepSeek V3.2
coding
评测
社区和从业者反馈在基准测试和价格之上增加了真实世界的信号。
Reviews
Be the first to review this model
Share your experience with DeepSeek V3.2 Exp and help the community make better decisions.
常见问题
关键信息
LMSYS Arena Elo
1422
Percentile
87.0
Weight
30%
No task benchmark data available yet for this model.
价格工具
成本估算器
每月比类别平均节省$39.55