Grok 4.20 0309 v2 vs Llama 3.1 Nemotron Ultra 253B v1
xAI vs NVIDIA — benchmarks, pricing, and capabilities side by side.
- •Grok 4.20 0309 v2 has the higher intelligence index (91.1 vs 78.3)
- •Llama 3.1 Nemotron Ultra 253B v1 is cheaper ($0.60 vs $2.00 per 1M input)
- •Grok 4.20 0309 v2 is faster
| Grok 4.20 0309 v2 | Llama 3.1 Nemotron Ultra 253B v1 | |
|---|---|---|
| Intelligence index | 91.1 | 78.3 |
| Developer | xAI | NVIDIA |
| Type | LLM | LLM |
| Access | — | Open weights |
| Context window | — | — |
| Input price | $2.00 / 1M | $0.60 / 1M |
| Output price | $6.00 / 1M | $1.80 / 1M |
| Speed | 105 tok/s | 42 tok/s |
| Released | April 7, 2026 | April 7, 2025 |
| Parameters | — | 253000000000 |
| Input modalities | — | — |
| Output modalities | — | — |
Shared benchmarks
Grok 4.20 0309 v2
Llama 3.1 Nemotron Ultra 253B v1
GPQA Diamond
91.1
76
Humanity’s Last Exam
32.2
8.1