Round score
0.6711
+67.11%
Prompts scored
27 / 32
median_e2e_speedup
Below tolerance
5
answered too little to score
Zeroed total
5
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 | +57.59% | 0.736s | 0.312s | 48 | — |
| hf-001 | +65.78% | 0.723s | 0.247s | 42 | — |
| hf-002 | +71.32% | 0.722s | 0.207s |
32 requests
| Request | TTFT | Total | Tokens | Slowest gap | Stream |
|---|---|---|---|---|---|
| hf-000 | 113.64ms | 0.312s | 48 | 71.81ms | — |
| hf-001 | 95.99ms | 0.282s | 52 | 43.74ms | — |
| hf-002 | 85.22ms | 0.207s |
verdict
evidence
min logprob
num prompts
mean logprob
coverage ratio
quantile logprob
num positions scored
role
| 42 |
| — |
| hf-003 | +63.78% | 0.915s | 0.331s | 48 | — |
| hf-004 | +56.87% | 0.782s | 0.337s | 43 | — |
| hf-005 | +72.90% | 0.785s | 0.213s | 42 | — |
| hf-006 | — | — | — | 27 | candidate output below tolerance |
| hf-007 | +77.58% | 0.783s | 0.176s | 42 | — |
| hf-008 | +60.20% | 0.784s | 0.312s | 42 | — |
| hf-009 | +73.66% | 0.775s | 0.204s | 43 | — |
| hf-011 | — | — | — | 42 | candidate output below tolerance |
| hf-010 | +72.61% | 0.738s | 0.202s | 42 | — |
| hf-012 | +72.45% | 0.763s | 0.210s | 43 | — |
| hf-015 | +55.89% | 0.342s | 0.151s | 16 | — |
| hf-013 | — | — | — | 28 | candidate output below tolerance |
| hf-014 | +78.75% | 0.901s | 0.191s | 46 | — |
| hf-016 | +75.23% | 0.768s | 0.190s | 43 | — |
| hf-018 | +74.34% | 0.741s | 0.190s | 40 | — |
| hf-017 | — | — | — | 43 | candidate output below tolerance |
| hf-021 | +72.80% | 0.760s | 0.207s | 40 | — |
| hf-019 | +76.06% | 1.326s | 0.317s | 73 | — |
| hf-020 | — | — | — | 23 | candidate output below tolerance |
| hf-022 | +72.48% | 0.780s | 0.215s | 42 | — |
| hf-023 | +62.35% | 0.740s | 0.279s | 39 | — |
| hf-024 | +68.45% | 0.810s | 0.255s | 45 | — |
| hf-025 | +73.93% | 0.780s | 0.203s | 43 | — |
| hf-027 | +71.49% | 0.538s | 0.154s | 28 | — |
| hf-026 | +74.31% | 0.771s | 0.198s | 42 | — |
| hf-029 | +54.08% | 0.710s | 0.326s | 43 | — |
| hf-028 | +64.92% | 0.946s | 0.332s | 59 | — |
| hf-030 | +51.72% | 0.615s | 0.297s | 41 | — |
| hf-031 | +52.14% | 0.550s | 0.263s | 43 | — |
| 45 |
| 15.56ms |
| — |
| hf-003 | 98.33ms | 0.331s | 48 | 118.55ms | — |
| hf-004 | 129.82ms | 0.337s | 44 | 70.47ms | — |
| hf-005 | 105.01ms | 0.213s | 42 | 16.25ms | — |
| hf-006 | 86.60ms | 0.178s | 27 | 16.03ms | — |
| hf-007 | 68.03ms | 0.176s | 42 | 16.17ms | — |
| hf-008 | 137.49ms | 0.312s | 42 | 40.97ms | — |
| hf-009 | 96.18ms | 0.204s | 43 | 16.04ms | — |
| hf-010 | 79.50ms | 0.202s | 42 | 16.07ms | — |
| hf-011 | 116.75ms | 0.289s | 42 | 44.38ms | — |
| hf-012 | 88.46ms | 0.210s | 43 | 15.86ms | — |
| hf-013 | 84.12ms | 0.191s | 28 | 16.78ms | — |
| hf-014 | 67.98ms | 0.191s | 46 | 18.02ms | — |
| hf-015 | 92.48ms | 0.166s | 20 | 15.76ms | — |
| hf-016 | 67.83ms | 0.190s | 43 | 15.86ms | — |
| hf-017 | 64.81ms | 0.187s | 43 | 16.26ms | — |
| hf-018 | 67.95ms | 0.190s | 42 | 15.92ms | — |
| hf-019 | 56.44ms | 0.317s | 73 | 43.62ms | — |
| hf-020 | 86.40ms | 0.164s | 23 | 15.71ms | — |
| hf-021 | 97.95ms | 0.207s | 40 | 16.01ms | — |
| hf-022 | 91.34ms | 0.215s | 42 | 16.23ms | — |
| hf-023 | 102.93ms | 0.279s | 39 | 67.72ms | — |
| hf-024 | 105.12ms | 0.255s | 45 | 42.76ms | — |
| hf-025 | 80.59ms | 0.203s | 43 | 15.98ms | — |
| hf-026 | 90.17ms | 0.198s | 42 | 16.28ms | — |
| hf-027 | 75.97ms | 0.199s | 44 | 16.29ms | — |
| hf-028 | 78.34ms | 0.347s | 61 | 70.86ms | — |
| hf-029 | 115.39ms | 0.326s | 43 | 75.90ms | — |
| hf-030 | 125.86ms | 0.297s | 41 | 49.40ms | — |
| hf-031 | 97.01ms | 0.263s | 43 | 15.58ms | — |
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