Round score
0.3178
+31.78%
Prompts scored
23 / 32
median_e2e_speedup
Below tolerance
9
answered too little to score
Zeroed total
9
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-001 | +31.92% | 0.396s | 0.269s | 18 | — |
| hf-000 | +49.77% | 0.740s | 0.372s | 40 | — |
| hf-002 | +27.69% | 0.528s | 0.382s |
32 requests
| Request | TTFT | Total | Tokens | Slowest gap | Stream |
|---|---|---|---|---|---|
| hf-000 | 155.32ms | 0.372s | 47 | 76.27ms | — |
| hf-001 | 144.29ms | 0.269s | 18 | 55.49ms | — |
| hf-002 | 163.54ms | 0.423s |
verdict
evidence
min logprob
num prompts
mean logprob
coverage ratio
quantile logprob
num positions scored
role
| 28 |
| — |
| hf-003 | — | — | — | 32 | candidate output below tolerance |
| hf-004 | +24.75% | 0.765s | 0.576s | 41 | — |
| hf-005 | — | — | — | 27 | candidate output below tolerance |
| hf-006 | — | — | — | 35 | candidate output below tolerance |
| hf-007 | +46.74% | 0.778s | 0.414s | 41 | — |
| hf-008 | +45.95% | 0.865s | 0.467s | 42 | — |
| hf-009 | +54.76% | 0.775s | 0.350s | 41 | — |
| hf-010 | — | — | — | 24 | candidate output below tolerance |
| hf-011 | +50.56% | 0.768s | 0.380s | 46 | — |
| hf-012 | +51.14% | 0.565s | 0.276s | 34 | — |
| hf-013 | +48.71% | 0.739s | 0.379s | 40 | — |
| hf-014 | +19.31% | 0.723s | 0.584s | 42 | — |
| hf-015 | +27.93% | 0.821s | 0.592s | 47 | — |
| hf-017 | -1.52% | 0.486s | 0.494s | 28 | — |
| hf-016 | +38.62% | 0.910s | 0.559s | 51 | — |
| hf-018 | — | — | — | 48 | candidate output below tolerance |
| hf-019 | +56.76% | 1.002s | 0.433s | 50 | — |
| hf-020 | +54.95% | 0.941s | 0.424s | 40 | — |
| hf-022 | +41.88% | 0.612s | 0.355s | 24 | — |
| hf-023 | +59.50% | 1.017s | 0.412s | 46 | — |
| hf-021 | +58.62% | 1.020s | 0.422s | 45 | — |
| hf-024 | +31.64% | 1.017s | 0.695s | 47 | — |
| hf-025 | +39.86% | 1.015s | 0.610s | 46 | — |
| hf-026 | — | — | — | 18 | candidate output below tolerance |
| hf-027 | — | — | — | 41 | candidate output below tolerance |
| hf-028 | -6.92% | 0.808s | 0.864s | 41 | — |
| hf-030 | — | — | — | 32 | candidate output below tolerance |
| hf-029 | — | — | — | 22 | candidate output below tolerance |
| hf-031 | +50.47% | 0.987s | 0.489s | 75 | — |
| 43 |
| 74.36ms |
| — |
| hf-003 | 145.44ms | 0.373s | 32 | 96.67ms | — |
| hf-004 | 177.74ms | 0.576s | 41 | 132.19ms | — |
| hf-005 | 229.79ms | 0.580s | 27 | 83.75ms | — |
| hf-006 | 143.01ms | 0.590s | 35 | 102.88ms | — |
| hf-007 | 150.59ms | 0.414s | 41 | 102.75ms | — |
| hf-008 | 189.78ms | 0.550s | 47 | 82.49ms | — |
| hf-009 | 146.08ms | 0.376s | 43 | 82.49ms | — |
| hf-010 | 157.07ms | 0.378s | 24 | 74.74ms | — |
| hf-011 | 152.76ms | 0.380s | 46 | 82.22ms | — |
| hf-012 | 154.82ms | 0.359s | 43 | 82.47ms | — |
| hf-013 | 158.32ms | 0.379s | 42 | 78.81ms | — |
| hf-014 | 154.57ms | 0.584s | 42 | 141.25ms | — |
| hf-015 | 233.42ms | 0.592s | 47 | 83.09ms | — |
| hf-016 | 180.56ms | 0.559s | 51 | 145.24ms | — |
| hf-017 | 164.73ms | 0.601s | 45 | 148.51ms | — |
| hf-018 | 233.38ms | 0.633s | 48 | 98.47ms | — |
| hf-019 | 174.37ms | 0.433s | 50 | 97.89ms | — |
| hf-020 | 182.75ms | 0.424s | 44 | 85.75ms | — |
| hf-021 | 173.10ms | 0.422s | 45 | 80.59ms | — |
| hf-022 | 153.60ms | 0.418s | 52 | 63.07ms | — |
| hf-023 | 133.94ms | 0.633s | 69 | 128.43ms | — |
| hf-024 | 213.25ms | 0.695s | 47 | 121.58ms | — |
| hf-025 | 212.40ms | 0.610s | 46 | 115.20ms | — |
| hf-026 | 153.02ms | 0.226s | 18 | 26.97ms | — |
| hf-027 | 207.63ms | 0.781s | 41 | 196.03ms | — |
| hf-028 | 202.51ms | 0.864s | 42 | 196.11ms | — |
| hf-029 | 381.17ms | 0.712s | 22 | 111.19ms | — |
| hf-030 | 292.53ms | 0.575s | 32 | 79.00ms | — |
| hf-031 | 234.46ms | 0.489s | 75 | 29.28ms | — |
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