Gemini 1.0 Ultra (Google DeepMind) and Grok 3 (xAI) are both frontier AI models. Gemini 1.0 Ultra was published in December 2023 and Grok 3 in February 2025.
Grok 3 was trained on 3.5 x 10^26 FLOP, about 7.0x the compute of Gemini 1.0 Ultra at 5 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.
These two models share no benchmark on which both have been scored, so no direct performance comparison is possible here. The specification table below is a comparison of inputs, not of results.
| Field | Gemini 1.0 Ultra | Grok 3 |
|---|---|---|
| Organization | Google DeepMind | xAI |
| Published | Dec 6, 2023 | Feb 17, 2025 |
| Training compute | 5 x 10^25 FLOP | 3.5 x 10^26 FLOP |
| Parameters | — | 3T |
| Training hardware | Google TPU v4 | NVIDIA H100 SXM5 80GB |
| Chips used | 57,000 | 80,000 |
| Training time | 2.4K h | 2.2K h |
| Training cost (2023 USD) | $31M | $218M |
| Training power draw | 38.4 MW | 109.9 MW |
| Accessibility | API access | API access |
| Open weights | No | No |
| Country | United States | United States |
