Qwen 3.8 Flash Next 176B-A6B — MI355X vs Vera Rubin NVL72 Performance per Dollar
Cost per million tokens of MI355X (AMD CDNA 4) versus Vera Rubin NVL72 (NVIDIA Vera Rubin) on Qwen 3.8 Flash Next 176B-A6B. Large-hyperscaler-volume ownership TCO normalized by total 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 →
Vera Rubin NVL72 edges MI355X at 86 tok/s/user on Qwen 3.8 Flash Next 176B-A6B — $0.00 per million tokens versus $0.01, a 78% cost-per-token gap.
Push Qwen 3.8 Flash Next 176B-A6B to 106 tok/s/user and MI355X lands at $0.01 per million tokens against Vera Rubin NVL72's $0.00 — Vera Rubin NVL72 pulls ahead by 167%.
MI355X: $0.02 per million tokens. Vera Rubin NVL72: $0.00. Both at 125 tok/s/user on Qwen 3.8 Flash Next 176B-A6B, with Vera Rubin NVL72 360% cheaper. (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): MI355X $1.50/chip/hr · Vera Rubin NVL72 $3.61/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 | MI355X:$0.007Vera Rubin NVL72:$0.004 | MI355X:$0.011Vera Rubin NVL72:$0.004 | MI355X:$0.021Vera Rubin NVL72:$0.005 |
| Concurrency | MI355X:~13Vera Rubin NVL72:~207 | MI355X:~8Vera Rubin NVL72:~234 | MI355X:~5Vera Rubin NVL72:~239 |
Inference Performance
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.