Round score
0.7187
+71.87%
Prompts scored
29 / 32
median_e2e_speedup
Below tolerance
3
answered too little to score
Zeroed total
3
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 | +71.26% | 0.733s | 0.211s | 42 | — |
| hf-001 | +73.32% | 0.684s | 0.182s | 40 | — |
| hf-002 | +77.69% | 0.687s | 0.153s |
32 requests
| Request | TTFT | Total | Tokens | Slowest gap | Stream |
|---|---|---|---|---|---|
| hf-000 | 98.87ms | 0.211s | 42 | 17.84ms | — |
| hf-001 | 102.09ms | 0.198s | 47 | 17.50ms | — |
| hf-002 | 72.46ms | 0.169s |
verdict
evidence
min logprob
num prompts
mean logprob
coverage ratio
quantile logprob
num positions scored
role
| 43 |
| — |
| hf-003 | +70.36% | 0.583s | 0.173s | 40 | — |
| hf-004 | +73.84% | 0.738s | 0.193s | 52 | — |
| hf-005 | +58.25% | 0.704s | 0.294s | 45 | — |
| hf-006 | +73.17% | 0.710s | 0.191s | 42 | — |
| hf-007 | +73.90% | 0.702s | 0.183s | 41 | — |
| hf-008 | +72.66% | 0.708s | 0.194s | 40 | — |
| hf-009 | +79.48% | 0.908s | 0.186s | 52 | — |
| hf-010 | — | — | — | 21 | candidate output below tolerance |
| hf-011 | +69.91% | 0.719s | 0.216s | 42 | — |
| hf-012 | +74.25% | 0.793s | 0.204s | 46 | — |
| hf-013 | +74.82% | 0.687s | 0.173s | 40 | — |
| hf-014 | +75.63% | 0.706s | 0.172s | 42 | — |
| hf-015 | +57.73% | 0.765s | 0.324s | 47 | — |
| hf-016 | +73.39% | 0.760s | 0.202s | 47 | — |
| hf-017 | +75.95% | 0.727s | 0.175s | 41 | — |
| hf-018 | +78.39% | 0.755s | 0.163s | 43 | — |
| hf-019 | +74.76% | 0.752s | 0.190s | 43 | — |
| hf-020 | +70.29% | 0.731s | 0.217s | 40 | — |
| hf-021 | +75.54% | 0.749s | 0.183s | 42 | — |
| hf-022 | +70.39% | 0.711s | 0.211s | 42 | — |
| hf-023 | — | — | — | 41 | candidate output below tolerance |
| hf-025 | +58.80% | 0.733s | 0.302s | 43 | — |
| hf-026 | +62.17% | 0.755s | 0.285s | 43 | — |
| hf-027 | +62.93% | 0.788s | 0.292s | 51 | — |
| hf-028 | +68.94% | 0.607s | 0.189s | 41 | — |
| hf-030 | +63.65% | 0.356s | 0.129s | 21 | — |
| hf-029 | +72.49% | 0.676s | 0.186s | 48 | — |
| hf-024 | +67.18% | 2.119s | 0.696s | 154 | — |
| hf-031 | — | — | — | 48 | candidate output below tolerance |
| 47 |
| 17.81ms |
| — |
| hf-003 | 76.09ms | 0.188s | 42 | 18.14ms | — |
| hf-004 | 65.43ms | 0.209s | 58 | 17.90ms | — |
| hf-005 | 78.88ms | 0.294s | 45 | 46.69ms | — |
| hf-006 | 95.99ms | 0.191s | 42 | 16.53ms | — |
| hf-007 | 101.76ms | 0.183s | 41 | 17.27ms | — |
| hf-008 | 66.42ms | 0.194s | 40 | 17.48ms | — |
| hf-009 | 60.27ms | 0.257s | 68 | 40.54ms | — |
| hf-010 | 89.13ms | 0.138s | 21 | 17.86ms | — |
| hf-011 | 104.12ms | 0.256s | 47 | 39.98ms | — |
| hf-012 | 92.62ms | 0.204s | 46 | 17.32ms | — |
| hf-013 | 76.63ms | 0.173s | 40 | 17.99ms | — |
| hf-014 | 75.30ms | 0.172s | 42 | 17.69ms | — |
| hf-015 | 131.74ms | 0.340s | 62 | 41.94ms | — |
| hf-016 | 90.14ms | 0.202s | 47 | 16.87ms | — |
| hf-017 | 79.27ms | 0.175s | 43 | 17.18ms | — |
| hf-018 | 67.53ms | 0.163s | 43 | 18.53ms | — |
| hf-019 | 92.59ms | 0.190s | 43 | 17.59ms | — |
| hf-020 | 121.04ms | 0.217s | 40 | 17.37ms | — |
| hf-021 | 87.33ms | 0.183s | 42 | 17.27ms | — |
| hf-022 | 83.57ms | 0.211s | 42 | 17.84ms | — |
| hf-023 | 104.89ms | 0.217s | 41 | 17.18ms | — |
| hf-024 | 86.39ms | 0.696s | 154 | 56.97ms | — |
| hf-025 | 89.52ms | 0.302s | 43 | 56.97ms | — |
| hf-026 | 107.58ms | 0.285s | 46 | 39.06ms | — |
| hf-027 | 85.30ms | 0.326s | 68 | 41.83ms | — |
| hf-028 | 92.08ms | 0.204s | 42 | 16.89ms | — |
| hf-029 | 74.62ms | 0.217s | 67 | 17.36ms | — |
| hf-030 | 80.64ms | 0.176s | 47 | 17.71ms | — |
| hf-031 | 62.88ms | 0.220s | 48 | 17.43ms | — |
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