MiniMax M3 428B — H200 vs MI300X Performance per Dollar
Cost per million tokens of H200 (NVIDIA Hopper) versus MI300X (AMD CDNA 3) on MiniMax M3 428B. Owning-hyperscaler TCO normalized by output tokens — 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 — same interactions as the main inference chart.
At 48 tok/s/user on MiniMax M3 428B, H200 costs $0.31 per million tokens; MI300X costs $0.51. H200 is 65% more cost-efficient at this operating point.
H200 edges MI300X at 82 tok/s/user on MiniMax M3 428B — $0.51 per million tokens versus $1.31, a 159% cost-per-token gap.
Push MiniMax M3 428B to 115 tok/s/user and H200 lands at $0.77 per million tokens against MI300X's $1.82 — H200 pulls ahead by 137%. (Numbers reflect the default 1k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
Chip pricing (owning hyperscaler): H200 $1.22/chip/hr · MI300X $0.95/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:$0.307MI300X:$0.509 | H200:$0.505MI300X:$1.307 | H200:$0.769MI300X:$1.819 |
| Concurrency | H200:~50MI300X:~45 | H200:~18MI300X:~10 | H200:~9MI300X:~5 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.