Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and MI300X (AMD CDNA 3) on MiniMax M2.5/M2.7. Latency, throughput, and cost across LLM workloads. Use the chart controls below to switch sequences, precisions, and metrics — same interactions as the main inference chart.
Throughput at 31 tok/s/user on MiniMax M2.5/M2.7: B200 hits 12194 tok/s/chip, MI300X hits 1563. Per-million costs land at $0.04 and $0.17 respectively. B200 is 328% cheaper per token; B200 delivers 680% more tok/s/chip.
B200 / MI300X on MiniMax M2.5/M2.7 at 49 tok/s/user: 8355 / 1396 tok/s/chip, $0.06 / $0.19 per million tokens. B200 is 229% cheaper per token; B200 delivers 498% more tok/s/chip.
Toward the upper edge of the 14–84 tok/s/user interactivity band, at 67 tok/s/user on MiniMax M2.5/M2.7: B200 runs 4668 tok/s/chip at $0.10/M tokens, MI300X runs 1067 at $0.25/M. B200 is 140% cheaper per token; B200 delivers 338% more tok/s/chip. (Numbers reflect the default 8k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:12194.1MI300X:1563.4 | B200:8354.6MI300X:1396.3 | B200:4668.0MI300X:1066.7 |
| Cost ($/M tok) | B200:$0.039MI300X:$0.169 | B200:$0.058MI300X:$0.189 | B200:$0.103MI300X:$0.247 |
| tok/s/MW | B200:7131024MI300X:1124722 | B200:4885728MI300X:1004511 | B200:2729802MI300X:767436 |
| Concurrency | B200:~1019MI300X:~18 | B200:~128MI300X:~6 | B200:~16MI300X:~4 |
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
TCO $/chip/hr: VR200: 3.61H100: 1.17H200: 1.22B200: 1.73B300: 2.26GB200: 1.86GB300: 2.31MI300X: 0.95MI325X: 1.1MI355X: 1.5RTX6000PRO: 0.68Jalapeño (Teacup): 1.47
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 • Drag to pan • Double-click to reset • Click a point to pin tooltip
MiniMax M2.5/2.7 230B 8K / 1K Total Tokens per $1 TCO vs. Interactivity