Kimi K2.5/K2.6 1T — B200 vs MI355X
Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and MI355X (AMD CDNA 4) on Kimi K2.5/K2.6 1T. 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.
Setting 54 tok/s/user as the target on Kimi K2.5/K2.6 1T, B200 produces 1082 tok/s/GPU ($0.50 per million tokens) and MI355X produces 632 ($0.66). B200 is 33% cheaper per token; B200 delivers 71% more tok/s/GPU.
At 74 tok/s/user interactivity on Kimi K2.5/K2.6 1T, B200 delivers 700 tok/s/GPU at $0.77 per million tokens; MI355X delivers 516 tok/s/GPU at $0.80. B200 is 3% cheaper per token; B200 delivers 36% more tok/s/GPU at this point.
B200 posts 480 tok/s/GPU for $1.12 per million tokens at 94 tok/s/user on Kimi K2.5/K2.6 1T; MI355X posts 396 tok/s/GPU for $1.04. MI355X is 7% cheaper per token; B200 delivers 21% more tok/s/GPU. (Numbers reflect the default 1k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | B200:1082.4MI355X:631.8 | B200:699.9MI355X:516.1 | B200:480.2MI355X:396.0 |
| Cost ($/M tok) | B200:$0.496MI355X:$0.658 | B200:$0.771MI355X:$0.797 | B200:$1.117MI355X:$1.040 |
| tok/s/MW | B200:498813MI355X:238416 | B200:322529MI355X:194767 | B200:221278MI355X:149436 |
| Concurrency | B200:~43MI355X:~24 | B200:~20MI355X:~14 | B200:~10MI355X:~9 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.