DeepSeek V4.1 Flash 552B — H100 vs MI300X
Head-to-head AI inference benchmark comparison of H100 (NVIDIA Hopper) and MI300X (AMD CDNA 3) on DeepSeek V4.1 Flash 552B. 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 →
Throughput at 68 tok/s/user on DeepSeek V4.1 Flash 552B: H100 hits 19780 tok/s/chip, MI300X hits 11924. Per-million costs land at $0.02 and $0.02 respectively. H100 is 35% cheaper per token; H100 delivers 66% more tok/s/chip.
H100 / MI300X on DeepSeek V4.1 Flash 552B at 113 tok/s/user: 17622 / 6980 tok/s/chip, $0.02 / $0.04 per million tokens. H100 is 105% cheaper per token; H100 delivers 152% more tok/s/chip.
Toward the upper edge of the 22–205 tok/s/user interactivity band, at 159 tok/s/user on DeepSeek V4.1 Flash 552B: H100 runs 9502 tok/s/chip at $0.03/M tokens, MI300X runs 4040 at $0.07/M. H100 is 91% cheaper per token; H100 delivers 135% more tok/s/chip. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | H100:19779.8MI300X:11923.9 | H100:17621.6MI300X:6980.3 | H100:9501.5MI300X:4040.4 |
| Cost ($/M tok) | H100:$0.016MI300X:$0.022 | H100:$0.018MI300X:$0.038 | H100:$0.034MI300X:$0.065 |
| tok/s/MW | H100:14437817MI300X:8578379 | H100:12862480MI300X:5021810 | H100:6935426MI300X:2906738 |
| Concurrency | H100:~24MI300X:~9 | H100:~21MI300X:~5 | H100:~11MI300X:~1 |
Inference Performance
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.