Llama 3.3 70B — MI300X vs MI355X
Head-to-head AI inference benchmark comparison of MI300X (AMD CDNA 3) and MI355X (AMD CDNA 4) on Llama 3.3 70B. 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.
MI300X / MI355X on Llama 3.3 70B at 53 tok/s/user: 1185 / 2248 tok/s/GPU, $0.26 / $0.18 per million tokens. MI355X is 43% cheaper per token; MI355X delivers 90% more tok/s/GPU.
Around the middle of the 34–112 tok/s/user interactivity band, at 73 tok/s/user on Llama 3.3 70B: MI300X runs 660 tok/s/GPU at $0.48/M tokens, MI355X runs 1600 at $0.25/M. MI355X is 91% cheaper per token; MI355X delivers 142% more tok/s/GPU.
Setting 93 tok/s/user as the target on Llama 3.3 70B, MI300X produces 373 tok/s/GPU ($0.83 per million tokens) and MI355X produces 872 ($0.46). MI355X is 79% cheaper per token; MI355X delivers 134% more tok/s/GPU. (Numbers reflect the default 1k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | MI300X:1185.2MI355X:2248.4 | MI300X:660.0MI355X:1599.8 | MI300X:372.9MI355X:871.8 |
| Cost ($/M tok) | MI300X:$0.257MI355X:$0.179 | MI300X:$0.479MI355X:$0.251 | MI300X:$0.832MI355X:$0.464 |
| tok/s/MW | MI300X:662142MI355X:848453 | MI300X:368695MI355X:603685 | MI300X:208350MI355X:328969 |
| Concurrency | MI300X:~50MI355X:~47 | MI300X:~32MI355X:~51 | MI300X:~17MI355X:~19 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.