Kimi K3 2.8T — B200 vs MI355X Performance per Dollar
Cost per million tokens of B200 (NVIDIA Blackwell) versus MI355X (AMD CDNA 4) on Kimi K3 2.8T. Owning-hyperscaler TCO normalized by output tokens — performance per dollar across LLM workloads. Pick the more cost-efficient SKU at every target interactivity level. 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 →
B200: $0.11 per million tokens. MI355X: $0.11. Both at 42 tok/s/user on Kimi K3 2.8T, with B200 2% cheaper.
Around the middle of the 19–112 tok/s/user interactivity band — at 66 tok/s/user — B200 runs $0.15 per million tokens on Kimi K3 2.8T while MI355X runs $0.16. B200 is the cheaper choice by 11%.
On Kimi K3 2.8T at 89 tok/s/user, the per-million math comes out to $0.18 for B200 and $0.23 for MI355X; B200 delivers 25% more output per dollar. (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.)
Chip pricing (owning hyperscaler): B200 $1.73/chip/hr · MI355X $1.50/chip/hr. Source: SemiAnalysis Market July 2026 Pricing Surveys & AI Cloud TCO Model.
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Dollar per Million Tokens | B200:$0.105MI355X:$0.108 | B200:$0.146MI355X:$0.163 | B200:$0.182MI355X:$0.227 |
| Concurrency | B200:~17MI355X:~8 | B200:~11MI355X:~5 | B200:~8MI355X:~2 |
Inference Performance
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.