DeepSeek-V4-Flash vs Llama 3.1 Nemotron Ultra 253B v1
DeepSeek vs NVIDIA — benchmarks, pricing, and capabilities side by side.
- •DeepSeek-V4-Flash has the higher intelligence index (89.4 vs 78.3)
- •DeepSeek-V4-Flash is cheaper ($0.10 vs $0.60 per 1M input)
- •DeepSeek-V4-Flash is faster
| DeepSeek-V4-Flash | Llama 3.1 Nemotron Ultra 253B v1 | |
|---|---|---|
| Intelligence index | 89.4 | 78.3 |
| Developer | DeepSeek | NVIDIA |
| Type | LLM | LLM |
| Access | Open weights | Open weights |
| Context window | 1,048,576 tokens | — |
| Input price | $0.10 / 1M | $0.60 / 1M |
| Output price | $0.20 / 1M | $1.80 / 1M |
| Speed | 109 tok/s | 42 tok/s |
| Released | April 24, 2026 | April 7, 2025 |
| Parameters | 284B (13B active) | 253000000000 |
| Input modalities | Text | — |
| Output modalities | Text | — |
Shared benchmarks
DeepSeek-V4-Flash
Llama 3.1 Nemotron Ultra 253B v1
GPQA Diamond
89.4
76
Humanity’s Last Exam
32.1
8.1