Round score
0.7874
+78.74%
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 | — | — | — | 42 | candidate output below tolerance |
| hf-001 | — | — | — | 28 | candidate output below tolerance |
| hf-002 | +78.51% | 0.898s | 0.193s |
32 requests
| Request | TTFT | Total | Tokens | Slowest gap | Stream |
|---|---|---|---|---|---|
| hf-000 | 49.27ms | 0.147s | 42 | 14.70ms | — |
| hf-001 | 44.20ms | 0.130s | 28 | 14.94ms | — |
| hf-002 | 38.98ms | 0.193s |
verdict
evidence
min logprob
num prompts
mean logprob
coverage ratio
quantile logprob
num positions scored
role
| 52 |
| — |
| hf-005 | +77.31% | 0.357s | 0.081s | 22 | — |
| hf-003 | — | — | — | 36 | candidate output below tolerance |
| hf-004 | +84.05% | 0.727s | 0.116s | 42 | — |
| hf-006 | +85.09% | 0.714s | 0.107s | 42 | — |
| hf-007 | +78.03% | 0.753s | 0.165s | 46 | — |
| hf-008 | +82.96% | 0.727s | 0.124s | 44 | — |
| hf-009 | +82.81% | 0.727s | 0.125s | 40 | — |
| hf-010 | +80.79% | 0.751s | 0.144s | 43 | — |
| hf-011 | +83.17% | 0.727s | 0.122s | 40 | — |
| hf-012 | +77.19% | 0.537s | 0.123s | 29 | — |
| hf-013 | +78.97% | 0.729s | 0.153s | 41 | — |
| hf-014 | +83.44% | 0.755s | 0.125s | 42 | — |
| hf-015 | +84.22% | 0.771s | 0.122s | 43 | — |
| hf-016 | +82.44% | 0.770s | 0.135s | 42 | — |
| hf-017 | +82.76% | 0.879s | 0.152s | 50 | — |
| hf-020 | +70.13% | 0.307s | 0.092s | 17 | — |
| hf-019 | +82.40% | 0.692s | 0.122s | 43 | — |
| hf-018 | +77.12% | 0.537s | 0.123s | 28 | — |
| hf-021 | +78.29% | 0.595s | 0.129s | 41 | — |
| hf-023 | +75.45% | 0.384s | 0.094s | 21 | — |
| hf-022 | +82.50% | 0.704s | 0.123s | 41 | — |
| hf-024 | +79.41% | 0.739s | 0.152s | 45 | — |
| hf-025 | — | — | — | 36 | candidate output below tolerance |
| hf-026 | +81.64% | 0.704s | 0.129s | 44 | — |
| hf-027 | +76.49% | 0.735s | 0.173s | 47 | — |
| hf-028 | +81.85% | 0.599s | 0.109s | 41 | — |
| hf-029 | +77.37% | 0.605s | 0.137s | 42 | — |
| hf-030 | +76.50% | 0.582s | 0.137s | 43 | — |
| hf-031 | — | — | — | 47 | candidate output below tolerance |
| 54 |
| 14.63ms |
| — |
| hf-003 | 45.61ms | 0.159s | 36 | 14.71ms | — |
| hf-004 | 43.84ms | 0.116s | 42 | 15.63ms | — |
| hf-005 | 38.85ms | 0.124s | 47 | 14.69ms | — |
| hf-006 | 36.94ms | 0.107s | 42 | 14.43ms | — |
| hf-007 | 40.50ms | 0.165s | 48 | 14.27ms | — |
| hf-008 | 41.17ms | 0.124s | 44 | 14.23ms | — |
| hf-009 | 39.67ms | 0.125s | 43 | 14.55ms | — |
| hf-010 | 44.72ms | 0.158s | 59 | 15.62ms | — |
| hf-011 | 38.97ms | 0.122s | 40 | 14.46ms | — |
| hf-012 | 38.67ms | 0.123s | 29 | 14.36ms | — |
| hf-013 | 44.60ms | 0.153s | 41 | 14.51ms | — |
| hf-014 | 40.65ms | 0.125s | 42 | 14.53ms | — |
| hf-015 | 38.41ms | 0.122s | 43 | 14.35ms | — |
| hf-016 | 38.34ms | 0.135s | 43 | 14.52ms | — |
| hf-017 | 39.35ms | 0.152s | 50 | 14.39ms | — |
| hf-018 | 39.60ms | 0.123s | 28 | 14.87ms | — |
| hf-019 | 38.87ms | 0.122s | 43 | 14.46ms | — |
| hf-020 | 37.39ms | 0.092s | 17 | 14.06ms | — |
| hf-021 | 44.83ms | 0.129s | 41 | 14.47ms | — |
| hf-022 | 39.02ms | 0.123s | 41 | 14.52ms | — |
| hf-023 | 39.09ms | 0.137s | 45 | 14.41ms | — |
| hf-024 | 40.65ms | 0.194s | 68 | 14.47ms | — |
| hf-025 | 41.37ms | 0.140s | 36 | 14.43ms | — |
| hf-026 | 44.76ms | 0.158s | 60 | 14.38ms | — |
| hf-027 | 46.34ms | 0.186s | 49 | 14.60ms | — |
| hf-028 | 39.63ms | 0.109s | 43 | 14.28ms | — |
| hf-029 | 39.01ms | 0.137s | 42 | 14.38ms | — |
| hf-030 | 38.94ms | 0.137s | 43 | 14.34ms | — |
| hf-031 | 39.82ms | 0.165s | 47 | 14.56ms | — |
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