meta-llama/Llama-3.1-8B-Instruct

vllm unknown on 8x NVIDIA H100 80GB HBM3 (fp16) · seed 0 · run 019e3b30-65a9-7861-80f9-dca6bb363757 · signed

Headline metrics

TTFT P50 (ms)	14.77
TTFT P99 (ms)	—
Throughput (tok/s)	—
$/M tokens	—
J/token	—
Power avg (W)	—
Power peak (W)	—
WER mean	—
J / audio s	—

All metrics

accuracy	0.9600
accuracy_p05	0.8400
accuracy_p50	1.00
accuracy_p95	1.00
drift_rate	0.2000
mean_drift_turn	4.00
n_ok	5.00
n_samples	5.00
ok_rate	1.00
persona_consistency_mean	0.9600
persona_consistency_p50	1.00
persona_consistency_p95	1.00
tokens_out_total	1,842
total_p50_ms	2,870
ttft_p50_ms	14.77

Provenance

Model revision	`unknown00`
Provider	vllm
Hardware fingerprint	`550474fc9132129654f5d20c316eaffec99a1f67c0cc0b4641e60efb46ea79e7`
Dataset	persona-consistency-mini · `ac864a018c1eb985c7f697a997162da20039e99fff11927ec3139fa4f2d4e9e3`
Timestamp	2026-05-18T13:04:48.041646+00:00
Envelope JSON	65908e52253d.json

Verify this result

bench verify /bench/leaderboard/envelopes/65908e52253d.json

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