Round score
0.7232
+72.32%
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 | +71.23% | 0.733s | 0.211s | 42 | — |
| hf-001 | +73.31% | 0.684s | 0.183s | 40 | — |
| hf-002 | +75.46% | 0.687s | 0.169s |
32 requests
| Request | TTFT | Total | Tokens | Slowest gap | Stream |
|---|---|---|---|---|---|
| hf-000 | 98.01ms | 0.211s | 42 | 17.28ms | — |
| hf-001 | 102.58ms | 0.198s | 47 | 16.55ms | — |
| hf-002 | 72.17ms | 0.169s |
verdict
evidence
min logprob
num prompts
mean logprob
coverage ratio
quantile logprob
num positions scored
role
| 43 |
| — |
| hf-003 | +70.41% | 0.583s | 0.172s | 40 | — |
| hf-004 | +74.12% | 0.738s | 0.191s | 52 | — |
| hf-005 | +73.36% | 0.704s | 0.188s | 45 | — |
| hf-006 | +75.93% | 0.710s | 0.171s | 42 | — |
| hf-007 | +73.63% | 0.702s | 0.185s | 41 | — |
| hf-008 | +73.65% | 0.730s | 0.192s | 42 | — |
| hf-009 | +83.16% | 0.908s | 0.153s | 52 | — |
| hf-010 | — | — | — | 21 | candidate output below tolerance |
| hf-011 | +69.99% | 0.719s | 0.216s | 42 | — |
| hf-012 | +74.17% | 0.793s | 0.205s | 46 | — |
| hf-013 | — | — | — | 28 | candidate output below tolerance |
| hf-014 | +75.87% | 0.706s | 0.170s | 42 | — |
| hf-015 | +57.73% | 0.765s | 0.324s | 47 | — |
| hf-016 | +75.49% | 0.760s | 0.186s | 47 | — |
| hf-017 | +77.41% | 0.727s | 0.164s | 41 | — |
| hf-018 | +76.58% | 0.766s | 0.179s | 44 | — |
| hf-019 | +72.87% | 0.752s | 0.204s | 43 | — |
| hf-020 | +64.46% | 0.731s | 0.260s | 40 | — |
| hf-021 | +72.76% | 0.749s | 0.204s | 42 | — |
| hf-022 | +70.63% | 0.711s | 0.209s | 42 | — |
| hf-023 | — | — | — | 41 | candidate output below tolerance |
| hf-025 | +75.34% | 0.733s | 0.181s | 43 | — |
| hf-026 | +62.20% | 0.755s | 0.285s | 43 | — |
| hf-027 | +63.38% | 0.788s | 0.289s | 51 | — |
| hf-028 | +69.01% | 0.607s | 0.188s | 41 | — |
| hf-030 | +68.28% | 0.356s | 0.113s | 21 | — |
| hf-029 | +71.88% | 0.676s | 0.190s | 48 | — |
| hf-024 | +66.69% | 2.098s | 0.699s | 152 | — |
| hf-031 | — | — | — | 48 | candidate output below tolerance |
| 47 |
| 17.77ms |
| — |
| hf-003 | 76.84ms | 0.187s | 42 | 17.75ms | — |
| hf-004 | 64.48ms | 0.207s | 58 | 18.12ms | — |
| hf-005 | 77.01ms | 0.188s | 45 | 17.38ms | — |
| hf-006 | 74.49ms | 0.171s | 42 | 17.97ms | — |
| hf-007 | 103.89ms | 0.185s | 41 | 17.13ms | — |
| hf-008 | 65.75ms | 0.192s | 42 | 17.31ms | — |
| hf-009 | 57.42ms | 0.200s | 67 | 16.71ms | — |
| hf-010 | 72.84ms | 0.121s | 21 | 17.30ms | — |
| hf-011 | 103.94ms | 0.259s | 47 | 42.94ms | — |
| hf-012 | 93.71ms | 0.205s | 46 | 17.21ms | — |
| hf-013 | 75.46ms | 0.141s | 28 | 17.36ms | — |
| hf-014 | 74.92ms | 0.170s | 42 | 17.07ms | — |
| hf-015 | 131.92ms | 0.324s | 47 | 41.69ms | — |
| hf-016 | 90.42ms | 0.186s | 47 | 16.54ms | — |
| hf-017 | 68.59ms | 0.164s | 41 | 16.93ms | — |
| hf-018 | 68.68ms | 0.179s | 47 | 17.05ms | — |
| hf-019 | 92.79ms | 0.204s | 43 | 17.32ms | — |
| hf-020 | 136.20ms | 0.260s | 40 | 41.97ms | — |
| hf-021 | 92.92ms | 0.204s | 42 | 17.04ms | — |
| hf-022 | 97.39ms | 0.209s | 42 | 17.42ms | — |
| hf-023 | 104.86ms | 0.217s | 41 | 17.53ms | — |
| hf-024 | 88.30ms | 0.699s | 152 | 52.80ms | — |
| hf-025 | 82.36ms | 0.181s | 44 | 16.81ms | — |
| hf-026 | 108.54ms | 0.285s | 43 | 43.14ms | — |
| hf-027 | 81.97ms | 0.289s | 51 | 41.56ms | — |
| hf-028 | 93.20ms | 0.188s | 41 | 16.46ms | — |
| hf-029 | 62.95ms | 0.190s | 48 | 17.37ms | — |
| hf-030 | 64.08ms | 0.161s | 47 | 17.35ms | — |
| hf-031 | 63.22ms | 0.207s | 48 | 17.72ms | — |
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