DeepSeek V3.2 Speciale vs Llama 3.1 Nemotron 70B Instruct
DeepSeek vs NVIDIA — benchmarks, pricing, and capabilities side by side.
- •DeepSeek V3.2 Speciale has the higher intelligence index (89.9 vs 43.7)
- •DeepSeek V3.2 Speciale is cheaper ($0.29 vs $1.20 per 1M input)
| DeepSeek V3.2 Speciale | Llama 3.1 Nemotron 70B Instruct | |
|---|---|---|
| Intelligence index | 89.9 | 43.7 |
| Developer | DeepSeek | NVIDIA |
| Type | LLM | LLM |
| Access | Open weights | Open weights |
| Context window | 163,840 tokens | — |
| Input price | $0.29 / 1M | $1.20 / 1M |
| Output price | $0.43 / 1M | $1.20 / 1M |
| Speed | — | 292 tok/s |
| Released | December 1, 2025 | October 1, 2024 |
| Parameters | — | 70000000000 |
| Input modalities | Text | — |
| Output modalities | Text | — |
Shared benchmarks
DeepSeek V3.2 Speciale
Llama 3.1 Nemotron 70B Instruct
AIME 2025
96.7
11
GPQA Diamond
87.1
46.5
Humanity’s Last Exam
26.1
4.6
LiveCodeBench
89.6
16.9
MMLU-Pro
86.3
69