Round score
0.8188
+81.88%
Prompts scored
28 / 32
median_e2e_speedup
Below tolerance
4
answered too little to score
Zeroed total
4
scored 0.0 without measuring
The round score is the scoring rule applied to the prompts below. Both engines are compared at the same output token count, so a candidate that stops early cannot buy speed by answering less: a prompt under the tolerance bar scores nothing and shows its gate reason.
32 prompts
| Prompt | Speedup | Baseline | Candidate | Tokens | Gate |
|---|---|---|---|---|---|
| hf-000 | +71.42% | 0.384s | 0.110s | 21 | — |
| hf-001 | +83.44% | 0.715s | 0.118s | 41 | — |
| hf-002 | +82.44% | 0.761s | 0.134s |
32 requests
| Request | TTFT | Total | Tokens | Slowest gap | Stream |
|---|---|---|---|---|---|
| hf-000 | 54.35ms | 0.152s | 44 | 14.10ms | — |
| hf-001 | 47.31ms | 0.118s | 41 | 15.19ms | — |
| hf-002 | 36.96ms | 0.134s |
verdict
evidence
min logprob
num prompts
mean logprob
coverage ratio
quantile logprob
num positions scored
role
| 45 |
| — |
| hf-003 | +81.60% | 0.782s | 0.144s | 45 | — |
| hf-004 | +82.15% | 0.754s | 0.134s | 43 | — |
| hf-005 | +83.05% | 0.737s | 0.125s | 42 | — |
| hf-006 | +81.42% | 0.730s | 0.136s | 42 | — |
| hf-007 | +82.85% | 0.759s | 0.130s | 44 | — |
| hf-008 | — | — | — | 15 | candidate output below tolerance |
| hf-009 | +83.53% | 0.730s | 0.120s | 41 | — |
| hf-010 | — | — | — | 42 | candidate output below tolerance |
| hf-011 | +78.00% | 0.783s | 0.172s | 46 | — |
| hf-012 | +86.16% | 0.951s | 0.132s | 50 | — |
| hf-013 | +84.06% | 0.749s | 0.119s | 40 | — |
| hf-014 | +83.25% | 0.800s | 0.134s | 45 | — |
| hf-017 | +82.61% | 0.551s | 0.096s | 28 | — |
| hf-015 | +80.89% | 0.949s | 0.181s | 51 | — |
| hf-016 | +86.47% | 0.781s | 0.106s | 42 | — |
| hf-018 | +82.36% | 0.760s | 0.134s | 41 | — |
| hf-019 | +80.79% | 0.769s | 0.148s | 42 | — |
| hf-020 | — | — | — | 17 | candidate output below tolerance |
| hf-021 | +81.03% | 0.770s | 0.146s | 42 | — |
| hf-022 | +83.72% | 0.591s | 0.096s | 32 | — |
| hf-023 | +83.84% | 0.508s | 0.082s | 24 | — |
| hf-024 | +80.33% | 0.700s | 0.138s | 40 | — |
| hf-025 | +84.74% | 0.785s | 0.120s | 47 | — |
| hf-026 | +81.32% | 0.792s | 0.148s | 47 | — |
| hf-027 | +83.24% | 0.737s | 0.124s | 42 | — |
| hf-028 | +81.18% | 0.787s | 0.148s | 43 | — |
| hf-029 | — | — | — | 28 | candidate output below tolerance |
| hf-030 | +79.11% | 0.675s | 0.141s | 47 | — |
| hf-031 | +79.24% | 0.578s | 0.120s | 46 | — |
| 45 |
| 14.24ms |
| — |
| hf-003 | 46.10ms | 0.144s | 45 | 14.27ms | — |
| hf-004 | 37.51ms | 0.134s | 43 | 14.15ms | — |
| hf-005 | 41.72ms | 0.125s | 42 | 14.03ms | — |
| hf-006 | 38.65ms | 0.136s | 42 | 14.41ms | — |
| hf-007 | 46.35ms | 0.130s | 44 | 14.22ms | — |
| hf-008 | 39.36ms | 0.081s | 15 | 14.82ms | — |
| hf-009 | 37.73ms | 0.120s | 41 | 14.36ms | — |
| hf-010 | 45.00ms | 0.117s | 42 | 15.54ms | — |
| hf-011 | 46.60ms | 0.187s | 53 | 14.59ms | — |
| hf-012 | 45.84ms | 0.132s | 50 | 15.17ms | — |
| hf-013 | 36.35ms | 0.119s | 42 | 14.35ms | — |
| hf-014 | 36.67ms | 0.134s | 45 | 14.74ms | — |
| hf-015 | 41.98ms | 0.181s | 51 | 14.28ms | — |
| hf-016 | 36.66ms | 0.106s | 42 | 14.48ms | — |
| hf-017 | 39.86ms | 0.096s | 28 | 14.46ms | — |
| hf-018 | 36.99ms | 0.134s | 41 | 14.34ms | — |
| hf-019 | 36.92ms | 0.148s | 42 | 14.41ms | — |
| hf-020 | 47.11ms | 0.117s | 17 | 14.53ms | — |
| hf-021 | 46.52ms | 0.146s | 42 | 14.87ms | — |
| hf-022 | 40.27ms | 0.139s | 57 | 14.30ms | — |
| hf-023 | 39.93ms | 0.138s | 49 | 14.42ms | — |
| hf-024 | 39.75ms | 0.138s | 40 | 14.37ms | — |
| hf-025 | 37.29ms | 0.120s | 47 | 14.12ms | — |
| hf-026 | 37.17ms | 0.148s | 47 | 14.38ms | — |
| hf-027 | 40.09ms | 0.124s | 42 | 14.23ms | — |
| hf-028 | 36.56ms | 0.148s | 45 | 14.80ms | — |
| hf-029 | 35.96ms | 0.091s | 28 | 14.03ms | — |
| hf-030 | 43.30ms | 0.141s | 47 | 14.29ms | — |
| hf-031 | 37.76ms | 0.120s | 46 | 14.68ms | — |
evidence
requests per s
sla goodput ratio
output tokens per s
p99 e2e ms rel range
p99 itl ms rel range
p99 ttft ms rel range
Image digest
Patch hash