MiniMax M3 428B β H200 vs MI325X Performance per Dollar
Cost per million tokens of H200 (NVIDIA Hopper) versus MI325X (AMD CDNA 3) on MiniMax M3 428B. Large-hyperscaler-volume ownership TCO normalized by total tokens β performance per dollar across LLM workloads. Pick the more cost-efficient SKU at every target interactivity level. 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
On MiniMax M3 428B at 78 tok/s/user, the per-million math comes out to $0.04 for H200 and $0.35 for MI325X; H200 delivers 831% more total tokens per dollar.
At 146 tok/s/user on MiniMax M3 428B, H200 costs $0.04 per million tokens; MI325X costs $0.35. H200 is 714% more cost-efficient at this operating point.
H200 edges MI325X at 214 tok/s/user on MiniMax M3 428B β $0.09 per million tokens versus $0.35, a 298% cost-per-token gap. (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.)
Chip pricing (owning hyperscaler): H200 $1.22/chip/hr Β· MI325X $1.10/chip/hr. Source: SemiAnalysis Market July 2026 Pricing Surveys & AI Cloud TCO Model.
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Dollar per Million Tokens | H200:$0.038MI325X:$0.354 | H200:$0.043MI325X:$0.354 | H200:$0.089MI325X:$0.354 |
| Concurrency | H200:~14MI325X:~1 | H200:~10MI325X:~1 | H200:~4MI325X:~1 |
Inference Performance
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.