Head-to-head AI inference benchmark comparison of H200 (NVIDIA Hopper) and MI355X (AMD CDNA 4) on Qwen 3.8 Flash Next 176B-A6B. 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.
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
MI355X hits 58839 tok/s/chip for $0.01 per million tokens at 86 tok/s/user on Qwen 3.8 Flash Next 176B-A6B. No H200 data at this operating point.
MI355X: 36602 tok/s/chip, $0.01 per million tokens at 106 tok/s/user on Qwen 3.8 Flash Next 176B-A6B. H200 is unmeasured here.
At 125 tok/s/user on Qwen 3.8 Flash Next 176B-A6B, MI355X delivers 19807 tok/s/chip at $0.02 per million tokens; H200 hasn't been benchmarked at this target. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | H200:βMI355X:58838.9 | H200:βMI355X:36602.3 | H200:βMI355X:19807.3 |
| Cost ($/M tok) | H200:βMI355X:$0.007 | H200:βMI355X:$0.011 | H200:βMI355X:$0.021 |
| tok/s/MW | H200:βMI355X:28152565 | H200:βMI355X:17513060 | H200:βMI355X:9477154 |
| Concurrency | H200:βMI355X:~13 | H200:βMI355X:~8 | H200:βMI355X:~5 |
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
