Head-to-head AI inference benchmark comparison of H100 (NVIDIA Hopper) and MI355X (AMD CDNA 4) on DeepSeek V4.1 Flash 552B. 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
Near the low end of the 67β322 tok/s/user interactivity band, at 131 tok/s/user on DeepSeek V4.1 Flash 552B: H100 runs 13721 tok/s/chip at $0.02/M tokens, MI355X runs 119222 at $0.00/M. MI355X is 578% cheaper per token; MI355X delivers 769% more tok/s/chip.
Setting 195 tok/s/user as the target on DeepSeek V4.1 Flash 552B, H100 produces 6493 tok/s/chip ($0.05 per million tokens) and MI355X produces 61508 ($0.01). MI355X is 639% cheaper per token; MI355X delivers 847% more tok/s/chip.
At 259 tok/s/user interactivity on DeepSeek V4.1 Flash 552B, H100 delivers 4222 tok/s/chip at $0.08 per million tokens; MI355X delivers 57290 tok/s/chip at $0.01. MI355X is 958% cheaper per token; MI355X delivers 1257% more tok/s/chip at this point. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | H100:13720.7MI355X:119222.0 | H100:6493.0MI355X:61507.9 | H100:4222.3MI355X:57289.9 |
| Cost ($/M tok) | H100:$0.024MI355X:$0.003 | H100:$0.050MI355X:$0.007 | H100:$0.077MI355X:$0.007 |
| tok/s/MW | H100:10015087MI355X:57043997 | H100:4739421MI355X:29429593 | H100:3081974MI355X:27411419 |
| Concurrency | H100:~16MI355X:~35 | H100:~7MI355X:~32 | H100:~4MI355X:~16 |
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
No measurements to plot for this selection. Review the benchmark controls above or adjust quick filters.
