Round score
0.1219
median speedup minus reliability deduction
Prompts scored
32 / 32
median_e2e_speedup
Below tolerance
0
answered too little to score
Zeroed total
0
scored 0.0 without measuring
Request interval
Thinking
Context limit
The same inputs and output limits apply to every engine. Actual concurrency depends on response duration and engine scheduling.
Median speedup
Failed requests
Deduction
Final score
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 | Aligned output | Gate |
|---|---|---|---|---|---|
| hf-008 | +12.12% | 108.791s | 95.606s | 5120 | — |
| hf-015 | +12.15% | 108.836s | 95.612s | 5120 | — |
| hf-004 | +12.19% | 108.828s | 95.562s |
32 requests
| Request | Input tokens | Output limit | TTFT | Total | Tokens | Slowest gap | Stream |
|---|---|---|---|---|---|---|---|
| hf-000 | 16,245 | 5,120 | 244.44ms | 95.642s | 5120 | 445.26ms | — |
| hf-001 | 7,623 | 5,120 | 258.35ms | 95.625s |
verdict
evidence
min logprob
num prompts
mean logprob
coverage ratio
quantile logprob
num positions scored
role
| 5120 |
| — |
| hf-006 | +12.18% | 108.795s | 95.539s | 5120 | — |
| hf-027 | +12.18% | 108.764s | 95.517s | 5120 | — |
| hf-005 | +12.22% | 108.839s | 95.540s | 5120 | — |
| hf-013 | +12.20% | 108.840s | 95.564s | 5120 | — |
| hf-026 | +12.21% | 108.799s | 95.512s | 5120 | — |
| hf-030 | +12.14% | 108.790s | 95.584s | 5120 | — |
| hf-021 | +12.19% | 108.785s | 95.521s | 5120 | — |
| hf-022 | +12.20% | 108.784s | 95.515s | 5120 | — |
| hf-014 | +12.10% | 108.784s | 95.619s | 5120 | — |
| hf-009 | +12.20% | 108.800s | 95.528s | 5120 | — |
| hf-012 | +12.20% | 108.820s | 95.547s | 5120 | — |
| hf-023 | +12.21% | 108.804s | 95.518s | 5120 | — |
| hf-018 | +12.25% | 108.858s | 95.527s | 5120 | — |
| hf-024 | +12.22% | 108.807s | 95.516s | 5120 | — |
| hf-000 | +12.10% | 108.811s | 95.642s | 5120 | — |
| hf-002 | +12.20% | 108.858s | 95.576s | 5120 | — |
| hf-001 | +12.12% | 108.809s | 95.625s | 5120 | — |
| hf-007 | +12.15% | 108.826s | 95.607s | 5120 | — |
| hf-028 | +12.22% | 108.800s | 95.508s | 5120 | — |
| hf-031 | +12.18% | 108.833s | 95.581s | 5120 | — |
| hf-010 | +12.13% | 108.830s | 95.632s | 5120 | — |
| hf-011 | +12.12% | 108.817s | 95.629s | 5120 | — |
| hf-003 | +12.19% | 108.811s | 95.544s | 5120 | — |
| hf-020 | +12.20% | 108.816s | 95.543s | 5120 | — |
| hf-025 | +12.19% | 108.767s | 95.509s | 5120 | — |
| hf-017 | +12.15% | 108.821s | 95.601s | 5120 | — |
| hf-019 | +12.19% | 108.780s | 95.522s | 5120 | — |
| hf-016 | +12.22% | 108.823s | 95.527s | 5120 | — |
| hf-029 | +12.20% | 108.822s | 95.544s | 5120 | — |
| 5120 |
| 429.61ms |
| — |
| hf-002 | 4,047 | 5,120 | 254.64ms | 95.576s | 5120 | 434.62ms | — |
| hf-003 | 16,210 | 5,120 | 254.59ms | 95.544s | 5120 | 428.57ms | — |
| hf-004 | 4,023 | 5,120 | 250.68ms | 95.562s | 5120 | 434.66ms | — |
| hf-005 | 31,898 | 5,120 | 326.02ms | 95.540s | 5120 | 352.94ms | — |
| hf-006 | 3,755 | 5,120 | 323.99ms | 95.539s | 5120 | 353.15ms | — |
| hf-007 | 16,052 | 5,120 | 378.96ms | 95.607s | 5120 | 296.12ms | — |
| hf-008 | 32,392 | 5,120 | 434.00ms | 95.606s | 5120 | 240.02ms | — |
| hf-009 | 8,018 | 5,120 | 375.14ms | 95.528s | 5120 | 295.53ms | — |
| hf-010 | 3,714 | 5,120 | 430.42ms | 95.632s | 5120 | 239.59ms | — |
| hf-011 | 15,971 | 5,120 | 428.50ms | 95.629s | 5120 | 239.16ms | — |
| hf-012 | 3,785 | 5,120 | 426.62ms | 95.547s | 5120 | 242.84ms | — |
| hf-013 | 16,163 | 5,120 | 446.10ms | 95.564s | 5120 | 221.27ms | — |
| hf-014 | 7,543 | 5,120 | 456.73ms | 95.619s | 5120 | 205.16ms | — |
| hf-015 | 29,768 | 5,120 | 479.99ms | 95.612s | 5120 | 179.67ms | — |
| hf-016 | 31,918 | 5,120 | 479.07ms | 95.527s | 5120 | 179.65ms | — |
| hf-017 | 29,730 | 5,120 | 510.51ms | 95.601s | 5120 | 146.40ms | — |
| hf-018 | 3,950 | 5,120 | 508.44ms | 95.527s | 5120 | 146.10ms | — |
| hf-019 | 15,843 | 5,120 | 506.25ms | 95.522s | 5120 | 147.27ms | — |
| hf-020 | 15,172 | 5,120 | 565.03ms | 95.543s | 5120 | 92.63ms | — |
| hf-021 | 32,739 | 5,120 | 563.57ms | 95.521s | 5120 | 84.80ms | — |
| hf-022 | 7,796 | 5,120 | 562.05ms | 95.515s | 5120 | 84.58ms | — |
| hf-023 | 8,092 | 5,120 | 559.88ms | 95.518s | 5120 | 84.78ms | — |
| hf-024 | 15,096 | 5,120 | 558.80ms | 95.516s | 5120 | 83.75ms | — |
| hf-025 | 3,924 | 5,120 | 557.40ms | 95.509s | 5120 | 83.34ms | — |
| hf-026 | 4,076 | 5,120 | 555.86ms | 95.512s | 5120 | 83.25ms | — |
| hf-027 | 7,625 | 5,120 | 555.02ms | 95.517s | 5120 | 120.77ms | — |
| hf-028 | 30,890 | 5,120 | 573.52ms | 95.508s | 5120 | 72.65ms | — |
| hf-029 | 32,394 | 5,120 | 612.06ms | 95.544s | 5120 | 152.56ms | — |
| hf-030 | 7,971 | 5,120 | 609.84ms | 95.584s | 5120 | 152.43ms | — |
| hf-031 | 7,683 | 5,120 | 607.62ms | 95.581s | 5120 | 152.33ms | — |
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