Head-to-head AI inference benchmark comparison of MI355X (AMD CDNA 4) and Vera Rubin NVL72 (NVIDIA Vera Rubin) on DeepSeekv4 Pro 0813 1.6T. 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.
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 →
Setting 102 tok/s/user as the target on DeepSeekv4 Pro 0813 1.6T, MI355X produces 7341 tok/s/chip ($0.06 per million tokens) and Vera Rubin NVL72 produces 194408 ($0.01). Vera Rubin NVL72 is 1000% cheaper per token; Vera Rubin NVL72 delivers 2548% more tok/s/chip.
At 131 tok/s/user interactivity on DeepSeekv4 Pro 0813 1.6T, MI355X delivers 4401 tok/s/chip at $0.09 per million tokens; Vera Rubin NVL72 delivers 161209 tok/s/chip at $0.01. Vera Rubin NVL72 is 1422% cheaper per token; Vera Rubin NVL72 delivers 3563% more tok/s/chip at this point.
MI355X posts 2664 tok/s/chip for $0.16 per million tokens at 159 tok/s/user on DeepSeekv4 Pro 0813 1.6T; Vera Rubin NVL72 posts 98612 tok/s/chip for $0.01. Vera Rubin NVL72 is 1438% cheaper per token; Vera Rubin NVL72 delivers 3602% more tok/s/chip. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | MI355X:7340.7Vera Rubin NVL72:194407.6 | MI355X:4400.8Vera Rubin NVL72:161208.7 | MI355X:2664.0Vera Rubin NVL72:98612.3 |
| Cost ($/M tok) | MI355X:$0.057Vera Rubin NVL72:$0.005 | MI355X:$0.095Vera Rubin NVL72:$0.006 | MI355X:$0.156Vera Rubin NVL72:$0.010 |
| tok/s/MW | MI355X:3512299Vera Rubin NVL72:58911397 | MI355X:2105647Vera Rubin NVL72:48851107 | MI355X:1274651Vera Rubin NVL72:29882512 |
| Concurrency | MI355X:~11Vera Rubin NVL72:~1747 | MI355X:~6Vera Rubin NVL72:~1058 | MI355X:~3Vera Rubin NVL72:~706 |
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 • Drag to pan • Double-click to reset • 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.
DeepSeek V4 Pro 0813 1.6T 8K / 1K Total Tokens per $1 TCO vs. Interactivity