Round score
0.6051
+60.51%
Prompts scored
26 / 32
median_e2e_speedup
Below tolerance
6
answered too little to score
Zeroed total
6
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 | +57.10% | 0.736s | 0.316s | 48 | — |
| hf-001 | +58.79% | 0.702s | 0.290s | 40 | — |
| hf-002 | +72.79% | 0.722s | 0.196s |
32 requests
| Request | TTFT | Total | Tokens | Slowest gap | Stream |
|---|---|---|---|---|---|
| hf-000 | 123.02ms | 0.316s | 48 | 95.55ms | — |
| hf-001 | 130.60ms | 0.290s | 40 | 69.26ms | — |
| hf-002 | 100.48ms | 0.196s |
verdict
evidence
min logprob
num prompts
mean logprob
coverage ratio
quantile logprob
num positions scored
role
| 42 |
| — |
| hf-003 | — | — | — | 41 | candidate output below tolerance |
| hf-004 | +57.85% | 0.782s | 0.330s | 43 | — |
| hf-005 | +64.79% | 0.785s | 0.276s | 42 | — |
| hf-006 | +76.83% | 0.902s | 0.209s | 45 | — |
| hf-007 | +74.18% | 0.783s | 0.202s | 42 | — |
| hf-008 | +53.90% | 0.784s | 0.362s | 42 | — |
| hf-009 | +62.08% | 0.775s | 0.294s | 43 | — |
| hf-011 | +59.85% | 0.771s | 0.310s | 47 | — |
| hf-010 | +72.05% | 0.738s | 0.206s | 42 | — |
| hf-012 | +73.11% | 0.763s | 0.205s | 43 | — |
| hf-015 | +50.85% | 0.342s | 0.168s | 16 | — |
| hf-013 | — | — | — | 28 | candidate output below tolerance |
| hf-014 | +78.36% | 0.914s | 0.198s | 48 | — |
| hf-016 | +72.10% | 0.768s | 0.214s | 43 | — |
| hf-018 | +72.78% | 0.741s | 0.202s | 40 | — |
| hf-017 | +71.04% | 0.943s | 0.273s | 53 | — |
| hf-021 | +71.85% | 0.782s | 0.220s | 42 | — |
| hf-019 | — | — | — | 21 | candidate output below tolerance |
| hf-020 | — | — | — | 23 | candidate output below tolerance |
| hf-022 | +60.12% | 0.780s | 0.311s | 42 | — |
| hf-023 | +58.19% | 0.762s | 0.319s | 41 | — |
| hf-024 | — | — | — | 31 | candidate output below tolerance |
| hf-025 | +72.48% | 0.747s | 0.206s | 40 | — |
| hf-027 | +67.14% | 0.538s | 0.177s | 28 | — |
| hf-026 | +60.91% | 0.760s | 0.297s | 41 | — |
| hf-029 | +71.01% | 0.710s | 0.206s | 43 | — |
| hf-028 | — | — | — | 47 | candidate output below tolerance |
| hf-030 | +48.90% | 0.615s | 0.315s | 41 | — |
| hf-031 | +59.03% | 0.550s | 0.225s | 43 | — |
| 42 |
| 16.56ms |
| — |
| hf-003 | 100.28ms | 0.313s | 41 | 100.31ms | — |
| hf-004 | 142.47ms | 0.330s | 43 | 93.93ms | — |
| hf-005 | 113.06ms | 0.276s | 42 | 66.48ms | — |
| hf-006 | 96.25ms | 0.209s | 45 | 16.96ms | — |
| hf-007 | 89.50ms | 0.202s | 42 | 17.15ms | — |
| hf-008 | 160.05ms | 0.362s | 42 | 67.53ms | — |
| hf-009 | 96.52ms | 0.294s | 43 | 68.67ms | — |
| hf-010 | 93.69ms | 0.206s | 42 | 17.01ms | — |
| hf-011 | 126.97ms | 0.310s | 47 | 69.29ms | — |
| hf-012 | 92.97ms | 0.205s | 43 | 16.78ms | — |
| hf-013 | 83.23ms | 0.179s | 28 | 16.77ms | — |
| hf-014 | 69.48ms | 0.198s | 48 | 16.89ms | — |
| hf-015 | 103.33ms | 0.283s | 42 | 67.36ms | — |
| hf-016 | 102.08ms | 0.214s | 43 | 16.51ms | — |
| hf-017 | 91.48ms | 0.273s | 53 | 69.15ms | — |
| hf-018 | 89.14ms | 0.202s | 40 | 16.96ms | — |
| hf-019 | 73.99ms | 0.123s | 21 | 16.77ms | — |
| hf-020 | 69.76ms | 0.151s | 23 | 16.60ms | — |
| hf-021 | 105.92ms | 0.220s | 42 | 17.00ms | — |
| hf-022 | 121.55ms | 0.311s | 42 | 97.89ms | — |
| hf-023 | 124.03ms | 0.319s | 41 | 102.93ms | — |
| hf-024 | 125.41ms | 0.333s | 31 | 67.76ms | — |
| hf-025 | 91.88ms | 0.206s | 40 | 16.92ms | — |
| hf-026 | 102.08ms | 0.297s | 41 | 67.96ms | — |
| hf-027 | 96.84ms | 0.177s | 28 | 16.55ms | — |
| hf-028 | 74.92ms | 0.203s | 47 | 16.92ms | — |
| hf-029 | 108.76ms | 0.302s | 47 | 95.94ms | — |
| hf-030 | 131.73ms | 0.315s | 41 | 68.17ms | — |
| hf-031 | 97.20ms | 0.241s | 50 | 18.03ms | — |
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