# Wed Oct 28 02:31:28 PM +03 2026 mode=batch server=127.0.0.1:18020
ROW 64conc 128in/512out | e2e=2496.15 tok/s | decode(C/medTPOT)=2696 | medTPOT=23.74 ms | meanTTFT=918.11 ms | dur=52.51s
ROW 64conc 256in/256out | e2e=1832.39 tok/s | decode(C/medTPOT)=2062 | medTPOT=31.04 ms | meanTTFT=1169.18 ms | dur=35.77s
ROW cohort C1 real prompts T=default | e2e=79.47 tok/s | decode(C/meanTPOT)=80.0 | tok/step=- | meanTTFT=82.08 ms
ROW cohort C2 real prompts T=default | e2e=146.05 tok/s | decode(C/meanTPOT)=154.9 | tok/step=- | meanTTFT=146.99 ms
ROW cohort C4 real prompts T=default | e2e=297.96 tok/s | decode(C/meanTPOT)=305.8 | tok/step=- | meanTTFT=150.17 ms
ROW cohort C8 real prompts T=default | e2e=546.51 tok/s | decode(C/meanTPOT)=565.4 | tok/step=- | meanTTFT=272.81 ms
# raw logs in /home/zeka/ai/benchmarks/hyperqwen-5090-2026-10-28/results/batch/run2
