Grok-4 Heavy vs Llama 3.2 90B Instruct
xAI vs Meta — benchmarks, pricing, and capabilities side by side.
- •Grok-4 Heavy has the higher intelligence index (89.3 vs 54)
| Grok-4 Heavy | Llama 3.2 90B Instruct | |
|---|---|---|
| Intelligence index | 89.3 | 54 |
| Developer | xAI | Meta |
| Type | Multimodal | Multimodal |
| Access | API only | Open weights |
| Context window | — | 128,000 tokens |
| Input price | — | $0.35 / 1M |
| Output price | — | $0.40 / 1M |
| Speed | — | 100 tok/s |
| Released | July 9, 2025 | September 25, 2024 |
| Parameters | — | 90000000000 |
| Input modalities | — | — |
| Output modalities | — | — |
Shared benchmarks
Grok-4 Heavy
Llama 3.2 90B Instruct
GPQA Diamond
88.4
46.7
Humanity’s Last Exam
50.7
4.9
LiveCodeBench
79.4
21.4