GPT-4 (Mar 2023) (OpenAI) and PaLM (540B) (Google Research) are both frontier AI models. GPT-4 (Mar 2023) was published in March 2023 and PaLM (540B) in April 2022.
GPT-4 (Mar 2023) was trained on 2.1 x 10^25 FLOP, about 8.3x the compute of PaLM (540B) at 2.5 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 | GPT-4 (Mar 2023) | PaLM (540B) |
|---|---|---|
| Organization | OpenAI | Google Research |
| Published | Mar 15, 2023 | Apr 4, 2022 |
| Training compute | 2.1 x 10^25 FLOP | 2.5 x 10^24 FLOP |
| Parameters | 1.8T | 540.4B |
| Dataset size | 5.4T | 780B |
| Training hardware | NVIDIA A100 SXM4 40 GB | Google TPU v4 |
| Chips used | 25,000 | 6,144 |
| Training time | 2.3K h | 1.5K h |
| Training cost (2023 USD) | $37M | $3M |
| Training power draw | 19.9 MW | 4.2 MW |
| Accessibility | API access | Unreleased |
| Open weights | No | No |
| Country | United States | United States |
