En vivo
Head to head

Flan-PaLM 540B vs Gemini 1.0 Ultra

Google
Flan-PaLM 540B
October 2022
vs
Google DeepMind
Gemini 1.0 Ultra
December 2023
2.5×10²⁴Training compute (FLOP)5×10²⁵
--Training cost$31M

Flan-PaLM 540B (Google) and Gemini 1.0 Ultra (Google DeepMind) are both frontier AI models. Flan-PaLM 540B was published in October 2022 and Gemini 1.0 Ultra in December 2023.

Gemini 1.0 Ultra was trained on 5×10²⁵ FLOP, about 19.7x the compute of Flan-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
Google
Organization
Google DeepMind
Oct 20, 2022
Published
Dec 6, 2023
2.5×10²⁴ FLOP
Training compute20x
5×10²⁵ FLOP
540B
Parameters
--
1.4B
Dataset size
--
Google TPU v4
Training hardware
Google TPU v4
512
Chips used111x
57,000
37 h
Training time
2,400 h
--
Training cost (2023 USD)
$31M
348.3 kW
Training power draw
38.4 MW
Unreleased
Accessibility
API access
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.