Head-to-head AI inference benchmark comparison of H100 (NVIDIA Hopper) and MI325X (AMD CDNA 3) on MiniMax M3 428B. 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
Setting 13 tok/s/user as the target on MiniMax M3 428B, H100 produces 654 tok/s/chip ($0.50 per million tokens) and MI325X produces 3310 ($0.09). MI325X is 439% cheaper per token; MI325X delivers 406% more tok/s/chip.
At 15 tok/s/user interactivity on MiniMax M3 428B, H100 delivers 637 tok/s/chip at $0.51 per million tokens; MI325X delivers 3255 tok/s/chip at $0.09. MI325X is 444% cheaper per token; MI325X delivers 411% more tok/s/chip at this point.
H100 posts 611 tok/s/chip for $0.53 per million tokens at 17 tok/s/user on MiniMax M3 428B; MI325X posts 2735 tok/s/chip for $0.11. MI325X is 376% cheaper per token; MI325X delivers 348% more tok/s/chip. (Numbers reflect the default agentic-traces Β· fp8 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | H100:653.7MI325X:3309.6 | H100:637.0MI325X:3255.5 | H100:610.7MI325X:2735.2 |
| Cost ($/M tok) | H100:$0.497MI325X:$0.092 | H100:$0.510MI325X:$0.094 | H100:$0.532MI325X:$0.112 |
| tok/s/MW | H100:477121MI325X:1958319 | H100:464987MI325X:1926302 | H100:445754MI325X:1618467 |
| Concurrency | H100:~2MI325X:~14 | H100:~1MI325X:~12 | H100:~1MI325X:~9 |
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
