Gemini 1.0 Ultra (Google DeepMind) and SenseChat (SenseTime) are both frontier AI models. Gemini 1.0 Ultra was published in December 2023 and SenseChat in April 2023.
Gemini 1.0 Ultra was trained on 5 x 10^25 FLOP, about 12.9x the compute of SenseChat at 3.9 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 | SenseChat |
|---|---|---|
| Organization | Google DeepMind | SenseTime |
| Published | Dec 6, 2023 | Apr 10, 2023 |
| Training compute | 5 x 10^25 FLOP | 3.9 x 10^24 FLOP |
| Parameters | — | 180B |
| Training hardware | Google TPU v4 | — |
| Chips used | 57,000 | — |
| Training time | 2.4K h | — |
| Training cost (2023 USD) | $31M | $5M |
| Training power draw | 38.4 MW | — |
| Accessibility | API access | API access |
| Open weights | No | No |
| Country | United States | Hong Kong |
