Home /
llm.inference.chatbot-short /
019e3b5c-f170-7b43-b9b9-0362535c3e25
microsoft/Phi-3.5-mini-instruct
vllm 0.21.0 on 8x NVIDIA H100 80GB HBM3
(fp16) · seed 42 · run 019e3b5c-f170-7b43-b9b9-0362535c3e25
· signed
Headline metrics
TTFT P50 (ms)
18.41
TTFT P99 (ms)
265
Throughput (tok/s)
716
$/M tokens
—
J/token
1.19
Power avg (W)
862
Power peak (W)
898
WER mean
—
J / audio s
—
All metrics
compliance_rate
0.9779
energy_joules_total
17,331
joules_per_token
1.19
ok_rate
1.00
power_avg_w
862
power_peak_w
898
req_per_s_all
6.67
req_per_s_passing
6.53
slo_hardware_class
h100
slo_template_resolved
ttft<200ms, tpot<50ms, total<3000ms
throughput_tok_per_s
716
total_p50_ms
531
total_p99_ms
787
tpot_p50_ms
4.66
tpot_p99_ms
6.57
ttft_p50_ms
18.41
ttft_p99_ms
265
Provenance
Model revision unknown00
Provider vllm
Hardware fingerprint 550474fc9132129654f5d20c316eaffec99a1f67c0cc0b4641e60efb46ea79e7
Dataset builtin-chatbot-short · 6f4b1f68fc3a813baa983cbe70cd9ef57f8c86e6b2e6ccc9aaa2a498e588d510
Timestamp 2026-05-18T13:53:27.408349+00:00
Envelope JSON 91eac8a2d0c4.json
Verify this result
bench verify /bench/leaderboard/envelopes/91eac8a2d0c4.json
bench verify re-downloads this envelope, recomputes the canonical
content hash, and validates the Sigstore signature against the embedded
certificate + Rekor entry.