实时
Head to head

Amazon Titan vs U-PaLM (540B)

Amazon
Amazon Titan
September 2023
vs
Google
U-PaLM (540B)
October 2022
4.8×10²⁴Training compute (FLOP)2.5×10²⁴
$8MTraining cost--

Amazon Titan (Amazon) and U-PaLM (540B) (Google) are both frontier AI models. Amazon Titan was published in September 2023 and U-PaLM (540B) in October 2022.

Amazon Titan was trained on 4.8×10²⁴ FLOP, about 1.9x the compute of U-PaLM (540B) at 2.5×10²⁴ 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.

Specifications
Amazon
Organization
Google
Sep 28, 2023
Published
Oct 20, 2022
4.8×10²⁴ FLOP
Training compute1.9x
2.5×10²⁴ FLOP
200B
Parameters2.7x
540B
4T
Dataset size3077x
1.3B
NVIDIA A100
Training hardware
Google TPU v4
13,760
Chips used27x
512
1,152 h
Training time
120 h
$8M
Training cost (2023 USD)
--
10.9 MW
Training power draw
348.3 kW
API access
Accessibility
Unreleased
No
Open weights
No
United States
Country
United States
Related comparisons
SourceEpoch AI, 'AI Models'. Published online at epoch.ai. Retrieved 2026-07-29 from https://epoch.ai/data/ai-models. Licensed under CC BY 4.0.