Round score
0.8393
+83.93%
Prompts scored
30 / 32
median_e2e_speedup
Below tolerance
2
answered too little to score
Zeroed total
2
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 | +75.09% | 0.384s | 0.096s | 21 | — |
| hf-001 | +85.07% | 0.715s | 0.107s | 41 | — |
| hf-002 | +82.28% | 0.761s | 0.135s |
32 requests
| Request | TTFT | Total | Tokens | Slowest gap | Stream |
|---|---|---|---|---|---|
| hf-000 | 38.31ms | 0.124s | 32 | 14.66ms | — |
| hf-001 | 31.71ms | 0.107s | 43 | 15.76ms | — |
| hf-002 | 32.28ms | 0.135s |
verdict
evidence
min logprob
num prompts
mean logprob
coverage ratio
quantile logprob
num positions scored
role
| 45 |
| — |
| hf-003 | +82.99% | 0.782s | 0.133s | 45 | — |
| hf-004 | +86.39% | 0.754s | 0.103s | 43 | — |
| hf-005 | +84.04% | 0.737s | 0.118s | 42 | — |
| hf-006 | +84.02% | 0.730s | 0.117s | 42 | — |
| hf-007 | +84.41% | 0.748s | 0.117s | 43 | — |
| hf-008 | +87.01% | 0.906s | 0.118s | 50 | — |
| hf-009 | +85.82% | 0.730s | 0.103s | 41 | — |
| hf-010 | — | — | — | 42 | candidate output below tolerance |
| hf-011 | +75.69% | 0.783s | 0.190s | 46 | — |
| hf-012 | +84.55% | 0.940s | 0.145s | 49 | — |
| hf-013 | +83.91% | 0.749s | 0.120s | 40 | — |
| hf-014 | +87.04% | 0.903s | 0.117s | 47 | — |
| hf-017 | +84.13% | 0.551s | 0.088s | 28 | — |
| hf-015 | +79.82% | 0.949s | 0.192s | 51 | — |
| hf-016 | +86.30% | 0.781s | 0.107s | 42 | — |
| hf-018 | +82.73% | 0.760s | 0.131s | 41 | — |
| hf-019 | +82.61% | 0.769s | 0.134s | 42 | — |
| hf-020 | +84.99% | 0.904s | 0.136s | 47 | — |
| hf-021 | +86.81% | 0.770s | 0.102s | 42 | — |
| hf-022 | +85.26% | 0.591s | 0.087s | 32 | — |
| hf-023 | +82.93% | 0.508s | 0.087s | 24 | — |
| hf-024 | +80.85% | 0.700s | 0.134s | 40 | — |
| hf-025 | +84.71% | 0.785s | 0.120s | 47 | — |
| hf-026 | +81.19% | 0.792s | 0.149s | 47 | — |
| hf-027 | +83.95% | 0.737s | 0.118s | 42 | — |
| hf-028 | +82.96% | 0.787s | 0.134s | 43 | — |
| hf-029 | +77.27% | 0.707s | 0.161s | 43 | — |
| hf-030 | +79.90% | 0.675s | 0.136s | 47 | — |
| hf-031 | — | — | — | 41 | candidate output below tolerance |
| 45 |
| 15.91ms |
| — |
| hf-003 | 33.49ms | 0.190s | 68 | 14.72ms | — |
| hf-004 | 31.73ms | 0.103s | 43 | 14.54ms | — |
| hf-005 | 32.57ms | 0.118s | 42 | 14.68ms | — |
| hf-006 | 31.72ms | 0.117s | 42 | 14.77ms | — |
| hf-007 | 30.79ms | 0.117s | 43 | 14.74ms | — |
| hf-008 | 32.84ms | 0.118s | 50 | 14.39ms | — |
| hf-009 | 32.90ms | 0.103s | 41 | 14.43ms | — |
| hf-010 | 31.25ms | 0.131s | 42 | 14.35ms | — |
| hf-011 | 33.20ms | 0.190s | 48 | 14.58ms | — |
| hf-012 | 30.50ms | 0.145s | 49 | 14.69ms | — |
| hf-013 | 32.42ms | 0.120s | 42 | 16.41ms | — |
| hf-014 | 31.39ms | 0.117s | 47 | 14.58ms | — |
| hf-015 | 30.95ms | 0.192s | 51 | 15.81ms | — |
| hf-016 | 32.27ms | 0.107s | 42 | 17.38ms | — |
| hf-017 | 30.80ms | 0.088s | 28 | 14.90ms | — |
| hf-018 | 31.20ms | 0.131s | 41 | 14.89ms | — |
| hf-019 | 30.36ms | 0.134s | 42 | 16.53ms | — |
| hf-020 | 31.75ms | 0.136s | 47 | 16.65ms | — |
| hf-021 | 30.74ms | 0.102s | 42 | 14.57ms | — |
| hf-022 | 30.76ms | 0.087s | 32 | 14.39ms | — |
| hf-023 | 30.16ms | 0.115s | 41 | 14.71ms | — |
| hf-024 | 30.79ms | 0.134s | 40 | 16.12ms | — |
| hf-025 | 30.73ms | 0.120s | 47 | 17.84ms | — |
| hf-026 | 31.24ms | 0.149s | 47 | 15.92ms | — |
| hf-027 | 33.67ms | 0.118s | 42 | 15.09ms | — |
| hf-028 | 31.02ms | 0.134s | 43 | 17.39ms | — |
| hf-029 | 31.80ms | 0.161s | 43 | 14.63ms | — |
| hf-030 | 32.24ms | 0.136s | 47 | 16.81ms | — |
| hf-031 | 31.84ms | 0.121s | 41 | 17.44ms | — |
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