RTX A6000 vLLM: Qwen/Qwen2.5-7B-Instruct median TTFT, 50 concurrent
583.54 ms (median time to first token, 50 concurrent)
Fields
id
benchmarks/dm-a6000-vllm-qwen25-7b-instruct-c50-ttftMetric
ttft_ms
Value
583.54
Unit
ms (median time to first token, 50 concurrent)
Model
qwen2.5-7b-instruct
Methodology
Database Mart vLLM bench on one RTX A6000 48 GB GDDR6 (page mislabels it 'Quadro RTX A6000'; the Ampere 8.6 / 10752-core spec block is the RTX A6000). Model Qwen/Qwen2.5-7B-Instruct, quantization column '16' (bf16/fp16 not distinguished; format_id null, no fp16 record exists). vLLM commit not printed. Single GPU, tensor parallel 1. 50 requests, input length 100 tokens, output length 600 tokens, benchmark duration 17.81 s, total input tokens 5000, total generated tokens 23817. AGGREGATE across the concurrent requests, not a per-request decode rate.
Measured by
third_party
Reproducible
yes
notes
Median TTFT for the 50-concurrent run.
Notes
Median TTFT for the 50-concurrent run.