Gemini 1.0 Ultra (Google DeepMind) and Minerva (540B) (Google) are both frontier AI models. Gemini 1.0 Ultra was published in December 2023 and Minerva (540B) in June 2022.
Gemini 1.0 Ultra was trained on 5 x 10^25 FLOP, about 18.2x the compute of Minerva (540B) at 2.7 x 10^24 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 | Minerva (540B) |
|---|---|---|
| Organization | Google DeepMind | |
| Published | Dec 6, 2023 | Jun 29, 2022 |
| Training compute | 5 x 10^25 FLOP | 2.7 x 10^24 FLOP |
| Parameters | — | 540.4B |
| Dataset size | — | 26B |
| Training hardware | Google TPU v4 | Google TPU v4 |
| Chips used | 57,000 | 1,024 |
| Training time | 2.4K h | 696 h |
| Training cost (2023 USD) | $31M | — |
| Training power draw | 38.4 MW | 698.4 kW |
| Accessibility | API access | Unreleased |
| Open weights | No | No |
| Country | United States | United States |
