Grok 4 (xAI) and Llama 3.1-405B (Meta AI) are both frontier AI models. Grok 4 was published in July 2025 and Llama 3.1-405B in July 2024.
Grok 4 was trained on 5 x 10^26 FLOP, about 13.2x the compute of Llama 3.1-405B at 3.8 x 10^25 FLOP. Training compute is the closest available proxy for how much was invested in a model, though it says nothing on its own about how well that compute was spent.
Grok 4 wins the single benchmark both were scored on. Benchmark counts are a crude scoreboard — the evaluations differ wildly in what they measure and in how saturated they are — so the per-benchmark table below matters more than the tally.
| Field | Grok 4 | Llama 3.1-405B |
|---|---|---|
| Organization | xAI | Meta AI |
| Published | Jul 9, 2025 | Jul 23, 2024 |
| Training compute | 5 x 10^26 FLOP | 3.8 x 10^25 FLOP |
| Parameters | 3T | 405B |
| Dataset size | — | 15.6T |
| Training hardware | — | NVIDIA H100 SXM5 80GB |
| Chips used | 200,000 | 16,384 |
| Training time | — | 2.1K h |
| Training cost (2023 USD) | $388M | $53M |
| Training power draw | — | 22.6 MW |
| Accessibility | API access | Open weights (restricted use) |
| Open weights | No | Yes |
| Country | United States | United States |
| Benchmark | Grok 4 | Llama 3.1-405B |
|---|---|---|
| Epoch Capabilities Index(ECI Score) | 147 | 129 |
