AI article
RTX 3090, RTX 5090, A100 and H100 on vLLM: the cheapest GPU flips with your answer time
Community description: I measured four GPUs serving the same 7B model on vLLM. Which one is cheapest per token has no single...
Dev.to | Nov 7, 2026 | Qaiser Mehdi
Automated excerpt
Under 3 s: H100 $0.162, RTX 5090 $0.382, A100 $0.803. Under 4 s: RTX 5090 $0.120, H100 $0.162, A100 $0.243. Under 8 s: RTX 5090 $0.050, H100 $0.099, A100 $0.106, RTX 3090 $0.490.
Selected automatically from source text; not independently written or fact-checked. Read the original for full context.