GPT-4.5 vs Llama 3.1-405B

GPT-4.5 (OpenAI) and Llama 3.1-405B (Meta AI) are both frontier AI models. GPT-4.5 was published in February 2025 and Llama 3.1-405B in July 2024.

GPT-4.5 was trained on 3.8×10²⁶ FLOP, about 10.0x the compute of Llama 3.1-405B at 3.8×10²⁵ 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.

GPT-4.5 wins the single benchmark both were scored on. Benchmark counts are a crude scoreboard — the evaluations differ wildly in what they measure and in how saturated they are — so the per-benchmark table below matters more than the tally.