DeepSeek V4.1 Flash 552B β GB200 NVL72 vs MI325X
Head-to-head AI inference benchmark comparison of GB200 NVL72 (NVIDIA Blackwell) and MI325X (AMD CDNA 3) on DeepSeek V4.1 Flash 552B. Latency, throughput, and cost across LLM workloads. Use the chart controls below to switch sequences, precisions, and metrics β same interactions as the main inference chart.
AgentX replays real coding-agent sessions rather than fixed-length prompts, so context grows turn over turn and most of each request is served from cache instead of being recomputed. That turns the comparison into a systems question: KV transfer between nodes, prefix-aware routing, and cache capacity all move the curve alongside raw chip throughput. Fixed-sequence workloads stay the clean baseline for kernel and silicon performance, so the two scenarios answer different questions about the same hardware. Learn more about AgentX β
At 117 tok/s/user interactivity on DeepSeek V4.1 Flash 552B, GB200 NVL72 delivers 352169 tok/s/chip at $0.00 per million tokens; MI325X delivers 7419 tok/s/chip at $0.04. GB200 NVL72 is 2707% cheaper per token; GB200 NVL72 delivers 4647% more tok/s/chip at this point.
GB200 NVL72 posts 282261 tok/s/chip for $0.00 per million tokens at 148 tok/s/user on DeepSeek V4.1 Flash 552B; MI325X posts 6496 tok/s/chip for $0.05. GB200 NVL72 is 2470% cheaper per token; GB200 NVL72 delivers 4245% more tok/s/chip.
Throughput at 179 tok/s/user on DeepSeek V4.1 Flash 552B: GB200 NVL72 hits 208760 tok/s/chip, MI325X hits 4148. Per-million costs land at $0.00 and $0.07 respectively. GB200 NVL72 is 2876% cheaper per token; GB200 NVL72 delivers 4933% more tok/s/chip. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | GB200 NVL72:352169.0MI325X:7418.6 | GB200 NVL72:282261.4MI325X:6496.4 | GB200 NVL72:208759.8MI325X:4148.0 |
| Cost ($/M tok) | GB200 NVL72:$0.001MI325X:$0.041 | GB200 NVL72:$0.002MI325X:$0.047 | GB200 NVL72:$0.002MI325X:$0.074 |
| tok/s/MW | GB200 NVL72:188325651MI325X:4389694 | GB200 NVL72:150941948MI325X:3844022 | GB200 NVL72:111636263MI325X:2454461 |
| Concurrency | GB200 NVL72:~248MI325X:~2 | GB200 NVL72:~188MI325X:~3 | GB200 NVL72:~138MI325X:~1 |
Inference Performance
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.