Head-to-head AI inference benchmark comparison of H100 (NVIDIA Hopper) and MI355X (AMD CDNA 4) on MiniMax M2.5/M2.7. 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.
H100 posts 1926 tok/s/chip for $0.17 per million tokens at 43 tok/s/user on MiniMax M2.5/M2.7; MI355X posts 4697 tok/s/chip for $0.09. MI355X is 90% cheaper per token; MI355X delivers 144% more tok/s/chip.
Throughput at 66 tok/s/user on MiniMax M2.5/M2.7: H100 hits 1269 tok/s/chip, MI355X hits 2917. Per-million costs land at $0.26 and $0.14 respectively. MI355X is 79% cheaper per token; MI355X delivers 130% more tok/s/chip.
H100 / MI355X on MiniMax M2.5/M2.7 at 88 tok/s/user: 823 / 1869 tok/s/chip, $0.39 / $0.22 per million tokens. MI355X is 77% cheaper per token; MI355X delivers 127% more tok/s/chip. (Numbers reflect the default 8k/1k Β· fp8 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | H100:1926.3MI355X:4697.0 | H100:1268.8MI355X:2917.5 | H100:823.2MI355X:1869.3 |
| Cost ($/M tok) | H100:$0.169MI355X:$0.089 | H100:$0.256MI355X:$0.143 | H100:$0.395MI355X:$0.223 |
| tok/s/MW | H100:1406025MI355X:2247388 | H100:926145MI355X:1395931 | H100:600902MI355X:894417 |
| Concurrency | H100:~41MI355X:~43 | H100:~18MI355X:~10 | H100:~9MI355X:~5 |
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
TCO $/chip/hr:
Source: SemiAnalysis Market July 2026 Pricing Surveys & AI Cloud TCO Model
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate cost per million tokens per decode chip or per prefill chip, rather than per total chip count. This makes direct token cost comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate input throughput per decode chip or per prefill chip, rather than per total chip count. This makes direct input throughput comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate output throughput per decode chip or per prefill chip, rather than per total chip count. This makes direct output throughput comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate power per decode chip or per prefill chip, rather than per total chip count. This makes direct power comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate Joules per decode chip or per prefill chip, rather than per total chip count. This makes direct Joules per token comparison with aggregated configs not an apples-to-apples comparison.
Matching measurements exist, but their chip series are hidden.
Shift+Scroll to zoom β’ Drag to pan β’ Double-click to reset β’ Click a point to pin tooltip
1 The ATOM engine is promising, however it has yet to serve production tokens. It is still in its infant stage.
MiniMax M2.5/2.7 230B 8K / 1K Total Tokens per $1 TCO vs. Interactivity