Round score
0.7275
+72.75%
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 | — | — | — | 42 | candidate output below tolerance |
| hf-002 | +64.45% | 0.726s | 0.258s | 42 | — |
| hf-003 | +72.38% | 0.706s | 0.195s |
32 requests
| Request | TTFT | Total | Tokens | Slowest gap | Stream |
|---|---|---|---|---|---|
| hf-000 | 98.89ms | 0.195s | 42 | 17.41ms | — |
| hf-001 | 126.08ms | 0.309s | 40 | 58.94ms | — |
| hf-002 | 109.51ms | 0.285s |
verdict
evidence
min logprob
num prompts
mean logprob
coverage ratio
quantile logprob
num positions scored
role
| 40 |
| — |
| hf-004 | +73.66% | 0.740s | 0.195s | 43 | — |
| hf-005 | +77.67% | 0.744s | 0.166s | 40 | — |
| hf-006 | +81.27% | 0.924s | 0.173s | 46 | — |
| hf-007 | +76.32% | 0.610s | 0.144s | 33 | — |
| hf-008 | +65.80% | 0.895s | 0.306s | 46 | — |
| hf-009 | +66.45% | 0.773s | 0.259s | 43 | — |
| hf-001 | — | — | — | 40 | candidate output below tolerance |
| hf-010 | +73.06% | 0.797s | 0.215s | 44 | — |
| hf-011 | +71.51% | 0.907s | 0.258s | 47 | — |
| hf-012 | +74.39% | 0.784s | 0.201s | 43 | — |
| hf-013 | +62.72% | 0.762s | 0.284s | 41 | — |
| hf-014 | +65.12% | 0.777s | 0.271s | 43 | — |
| hf-016 | +76.03% | 0.882s | 0.211s | 49 | — |
| hf-017 | — | — | — | 17 | candidate output below tolerance |
| hf-015 | +74.48% | 1.194s | 0.305s | 72 | — |
| hf-018 | +73.68% | 0.782s | 0.206s | 47 | — |
| hf-019 | +71.69% | 0.697s | 0.197s | 42 | — |
| hf-020 | +72.47% | 0.661s | 0.182s | 42 | — |
| hf-021 | +73.73% | 0.723s | 0.190s | 45 | — |
| hf-022 | +64.53% | 0.608s | 0.216s | 41 | — |
| hf-023 | +73.74% | 0.747s | 0.196s | 44 | — |
| hf-024 | +71.60% | 0.743s | 0.211s | 43 | — |
| hf-025 | +76.02% | 0.684s | 0.164s | 42 | — |
| hf-026 | +66.36% | 0.601s | 0.202s | 41 | — |
| hf-028 | +75.03% | 0.517s | 0.129s | 29 | — |
| hf-027 | +73.55% | 0.705s | 0.186s | 43 | — |
| hf-029 | +77.48% | 0.790s | 0.178s | 51 | — |
| hf-030 | +73.03% | 0.664s | 0.179s | 47 | — |
| hf-031 | +65.89% | 0.530s | 0.181s | 41 | — |
| 43 |
| 53.31ms |
| — |
| hf-003 | 85.16ms | 0.195s | 40 | 16.73ms | — |
| hf-004 | 69.34ms | 0.195s | 47 | 17.14ms | — |
| hf-005 | 70.46ms | 0.295s | 69 | 43.43ms | — |
| hf-006 | 94.79ms | 0.189s | 51 | 16.43ms | — |
| hf-007 | 62.92ms | 0.207s | 67 | 18.70ms | — |
| hf-008 | 142.76ms | 0.306s | 46 | 46.02ms | — |
| hf-009 | 106.06ms | 0.259s | 43 | 42.80ms | — |
| hf-010 | 119.41ms | 0.215s | 44 | 17.16ms | — |
| hf-011 | 104.27ms | 0.258s | 47 | 42.57ms | — |
| hf-012 | 89.43ms | 0.201s | 43 | 17.50ms | — |
| hf-013 | 93.41ms | 0.284s | 42 | 79.67ms | — |
| hf-014 | 122.97ms | 0.271s | 43 | 43.62ms | — |
| hf-015 | 70.82ms | 0.305s | 72 | 44.41ms | — |
| hf-016 | 86.03ms | 0.254s | 53 | 42.14ms | — |
| hf-017 | 89.29ms | 0.138s | 17 | 17.40ms | — |
| hf-018 | 63.32ms | 0.206s | 48 | 17.49ms | — |
| hf-019 | 85.61ms | 0.197s | 42 | 18.15ms | — |
| hf-020 | 71.22ms | 0.182s | 42 | 18.12ms | — |
| hf-021 | 63.46ms | 0.190s | 49 | 22.37ms | — |
| hf-022 | 102.67ms | 0.216s | 41 | 17.44ms | — |
| hf-023 | 85.63ms | 0.212s | 48 | 17.37ms | — |
| hf-024 | 84.19ms | 0.211s | 44 | 17.61ms | — |
| hf-025 | 83.67ms | 0.164s | 42 | 17.75ms | — |
| hf-026 | 91.44ms | 0.202s | 41 | 16.84ms | — |
| hf-027 | 75.53ms | 0.186s | 43 | 17.89ms | — |
| hf-028 | 63.97ms | 0.129s | 29 | 17.84ms | — |
| hf-029 | 66.15ms | 0.178s | 51 | 17.58ms | — |
| hf-030 | 84.36ms | 0.179s | 47 | 17.13ms | — |
| hf-031 | 85.13ms | 0.181s | 42 | 17.42ms | — |
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