Level1Techs: 4x RTX 3090, vLLM, Qwen3.5-27B, 50 concurrent reqs ~1024 tok ctx: ~348 tok/s aggregate output
348.0 tok/s output, aggregate over 50 concurrent requests
Fields
id
benchmarks/l1t-4x-rtx-3090-qwen35-27b-vllm-50req-tpsMetric
tps_aggregate
Value
348
Unit
tok/s output, aggregate over 50 concurrent requests
Model
Qwen3.5-27B
Methodology
FORUM MEASUREMENT (Level1Techs, relayed by hardware-corner.net B70 launch article). 4x RTX 3090 vLLM, Qwen3.5-27B, 50 concurrent requests, ~1024-token context: sustained ~348 tok/s output, avg TTFT ~18.7 s. AGGREGATE across 50 requests, not per-request.
Measured by
third_party
Reproducible
no
notes
Same-test baseline for the 4x B70 row: B70 edges throughput and roughly halves TTFT at this concurrency.
Notes
Same-test baseline for the 4x B70 row: B70 edges throughput and roughly halves TTFT at this concurrency.