Head-to-head AI inference benchmark comparison of MI325X (AMD CDNA 3) 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
MI325X / MI355X on DeepSeek V4.1 Flash 552B at 85 tok/s/user: 13027 / 129881 tok/s/chip, $0.02 / $0.00 per million tokens. MI355X is 631% cheaper per token; MI355X delivers 897% more tok/s/chip.
Around the middle of the 43β215 tok/s/user interactivity band, at 129 tok/s/user on DeepSeek V4.1 Flash 552B: MI325X runs 6911 tok/s/chip at $0.04/M tokens, MI355X runs 101582 at $0.00/M. MI355X is 978% cheaper per token; MI355X delivers 1370% more tok/s/chip.
Setting 172 tok/s/user as the target on DeepSeek V4.1 Flash 552B, MI325X produces 4081 tok/s/chip ($0.07 per million tokens) and MI355X produces 78779 ($0.01). MI355X is 1316% cheaper per token; MI355X delivers 1830% more tok/s/chip. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | MI325X:13027.2MI355X:129880.7 | MI325X:6910.8MI355X:101582.2 | MI325X:4081.1MI355X:78779.0 |
| Cost ($/M tok) | MI325X:$0.023MI355X:$0.003 | MI325X:$0.044MI355X:$0.004 | MI325X:$0.075MI355X:$0.005 |
| tok/s/MW | MI325X:7708385MI355X:62143866 | MI325X:4089236MI355X:48603928 | MI325X:2414867MI355X:37693284 |
| Concurrency | MI325X:~5MI355X:~58 | MI325X:~1MI355X:~32 | MI325X:~1MI355X:~32 |
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.
