DeepSeek V3.2 Speciale vs Qwen2.5 Max
DeepSeek vs Alibaba — benchmarks, pricing, and capabilities side by side.
- •DeepSeek V3.2 Speciale has the higher intelligence index (89.9 vs 63.6)
- •DeepSeek V3.2 Speciale is cheaper ($0.29 vs $1.60 per 1M input)
| DeepSeek V3.2 Speciale | Qwen2.5 Max | |
|---|---|---|
| Intelligence index | 89.9 | 63.6 |
| Developer | DeepSeek | Alibaba |
| Type | LLM | LLM |
| Access | Open weights | — |
| Context window | 163,840 tokens | — |
| Input price | $0.29 / 1M | $1.60 / 1M |
| Output price | $0.43 / 1M | $6.40 / 1M |
| Speed | — | 50 tok/s |
| Released | December 1, 2025 | January 28, 2025 |
| Parameters | — | — |
| Input modalities | Text | — |
| Output modalities | Text | — |
Shared benchmarks
DeepSeek V3.2 Speciale
Qwen2.5 Max
GPQA Diamond
87.1
58.7
Humanity’s Last Exam
26.1
4.5
LiveCodeBench
89.6
35.9
MMLU-Pro
86.3
76.2