Head-to-head AI inference benchmark comparison of B300 (NVIDIA Blackwell) and RTX PRO 6000 (NVIDIA Blackwell) on Qwen 3.5 397B-A17B. 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 →
B300 hits 45488 tok/s/chip for $0.01 per million tokens at 76 tok/s/user on Qwen 3.5 397B-A17B. No RTX PRO 6000 data at this operating point.
B300: 31100 tok/s/chip, $0.02 per million tokens at 145 tok/s/user on Qwen 3.5 397B-A17B. RTX PRO 6000 is unmeasured here.
At 213 tok/s/user on Qwen 3.5 397B-A17B, B300 delivers 16709 tok/s/chip at $0.04 per million tokens; RTX PRO 6000 hasn't been benchmarked at this target. (Numbers reflect the default agentic-traces · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B300:45487.9RTX PRO 6000:— | B300:31099.8RTX PRO 6000:— | B300:16708.5RTX PRO 6000:— |
| Cost ($/M tok) | B300:$0.014RTX PRO 6000:— | B300:$0.020RTX PRO 6000:— | B300:$0.038RTX PRO 6000:— |
| tok/s/MW | B300:23940979RTX PRO 6000:— | B300:16368341RTX PRO 6000:— | B300:8793955RTX PRO 6000:— |
| Concurrency | B300:~40RTX PRO 6000:— | B300:~22RTX PRO 6000:— | B300:~11RTX PRO 6000:— |
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
TCO $/chip/hr:
Source: SemiAnalysis Market July 2026 Pricing Surveys & AI Cloud TCO Model
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate cost per million tokens per decode chip or per prefill chip, rather than per total chip count. This makes direct token cost comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate input throughput per decode chip or per prefill chip, rather than per total chip count. This makes direct input throughput comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate output throughput per decode chip or per prefill chip, rather than per total chip count. This makes direct output throughput comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate power per decode chip or per prefill chip, rather than per total chip count. This makes direct power comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate Joules per decode chip or per prefill chip, rather than per total chip count. This makes direct Joules per token comparison with aggregated configs not an apples-to-apples comparison.
Matching measurements exist, but their chip series are hidden.
Shift+Scroll to zoom • Drag to pan • Double-click to reset • Click a point to pin tooltip
Qwen3.5 397B 8K / 1K Total Tokens per $1 TCO vs. Interactivity