InferenceBench

google/gemma-2-9b-it

vllm unknown on 8x NVIDIA H100 80GB HBM3 (fp16) · seed 0 · run 019e3b97-89b8-78d1-942f-d5fbdfe20496 · signed

Headline metrics

TTFT P50 (ms) 16.77
TTFT P99 (ms)
Throughput (tok/s)
$/M tokens
J/token
Power avg (W)
Power peak (W)
WER mean
J / audio s

All metrics

chrf_mean 0.7462
chrf_p50 0.6872
chrf_p95 1.00
n_ok 8.00
n_samples 8.00
ok_rate 1.00
tokens_out_total 199
total_p50_ms 292
ttft_p50_ms 16.77

Provenance

Model revisionunknown00
Providervllm
Hardware fingerprint550474fc9132129654f5d20c316eaffec99a1f67c0cc0b4641e60efb46ea79e7
Datasetbuiltin-flores-mini-en-fr · 25cb8c32504c769c88e3b41ad01bb304ccd100dbc9278b3a41f520053512185e
Timestamp2026-05-18T14:57:27.480069+00:00
Envelope JSONd09f9fe94ca2.json

Verify this result

bench verify /bench/leaderboard/envelopes/d09f9fe94ca2.json

bench verify re-downloads this envelope, recomputes the canonical content hash, and validates the Sigstore signature against the embedded certificate + Rekor entry.