Ao vivo
Head to head

Llama 4 Behemoth (preview) vs Llama Nemotron Ultra 253B

vs
5.2×10²⁵Training compute (FLOP)3.9×10²⁵
$45MTraining cost--

Llama 4 Behemoth (preview) (Meta AI) and Llama Nemotron Ultra 253B (NVIDIA) are both frontier AI models. Llama 4 Behemoth (preview) was published in April 2025 and Llama Nemotron Ultra 253B in March 2025.

Llama 4 Behemoth (preview) was trained on 5.2×10²⁵ FLOP, about 1.3x the compute of Llama Nemotron Ultra 253B at 3.9×10²⁵ 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
Meta AI
Organization
NVIDIA
Apr 5, 2025
Published
Mar 18, 2025
5.2×10²⁵ FLOP
Training compute1.3x
3.9×10²⁵ FLOP
2T
Parameters7.9x
253B
30T
Dataset size50x
603B
NVIDIA H100 SXM5 80GB
Training hardware
--
32,000
Chips used
--
$45M
Training cost (2023 USD)
--
43.9 MW
Training power draw
--
Unreleased
Accessibility
Open weights (restricted use)
No
Open weights
Yes
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.