Round score
0.7605
+76.05%
Prompts scored
25 / 32
median_e2e_speedup
Below tolerance
7
answered too little to score
Zeroed total
7
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 | — | — | — | 6 | candidate output below tolerance |
| hf-001 | +80.49% | 0.760s | 0.148s | 45 | — |
| hf-003 | +81.39% | 0.755s | 0.141s |
32 requests
| Request | TTFT | Total | Tokens | Slowest gap | Stream |
|---|---|---|---|---|---|
| hf-000 | 49.05ms | 0.094s | 6 | 15.36ms | — |
| hf-001 | 57.75ms | 0.148s | 45 | 15.49ms | — |
| hf-002 | 66.46ms | 0.113s |
verdict
evidence
min logprob
num prompts
mean logprob
coverage ratio
quantile logprob
num positions scored
role
| 44 |
| — |
| hf-002 | — | — | — | 21 | candidate output below tolerance |
| hf-004 | +83.13% | 0.730s | 0.123s | 42 | — |
| hf-005 | — | — | — | 20 | candidate output below tolerance |
| hf-006 | +77.41% | 0.694s | 0.157s | 42 | — |
| hf-007 | +75.97% | 0.671s | 0.161s | 43 | — |
| hf-009 | — | — | — | 28 | candidate output below tolerance |
| hf-008 | — | — | — | 51 | candidate output below tolerance |
| hf-010 | +77.69% | 0.656s | 0.146s | 41 | — |
| hf-012 | — | — | — | 23 | candidate output below tolerance |
| hf-011 | +78.54% | 0.903s | 0.194s | 52 | — |
| hf-013 | +68.36% | 0.497s | 0.157s | 28 | — |
| hf-014 | +83.94% | 0.739s | 0.119s | 42 | — |
| hf-015 | +65.19% | 0.717s | 0.250s | 41 | — |
| hf-016 | +79.96% | 0.739s | 0.148s | 43 | — |
| hf-017 | +76.13% | 0.742s | 0.177s | 43 | — |
| hf-018 | +76.85% | 0.704s | 0.163s | 42 | — |
| hf-019 | +77.80% | 0.872s | 0.194s | 54 | — |
| hf-020 | +78.59% | 0.662s | 0.142s | 41 | — |
| hf-021 | +77.54% | 0.717s | 0.161s | 47 | — |
| hf-023 | +74.90% | 0.764s | 0.192s | 52 | — |
| hf-022 | +81.75% | 0.687s | 0.125s | 45 | — |
| hf-024 | +76.39% | 0.701s | 0.166s | 46 | — |
| hf-026 | +55.90% | 0.359s | 0.158s | 23 | — |
| hf-025 | +78.55% | 0.591s | 0.127s | 42 | — |
| hf-029 | +63.61% | 0.325s | 0.118s | 20 | — |
| hf-028 | +74.69% | 0.708s | 0.179s | 41 | — |
| hf-027 | — | — | — | 38 | candidate output below tolerance |
| hf-030 | +74.33% | 0.586s | 0.151s | 41 | — |
| hf-031 | +70.99% | 0.297s | 0.086s | 20 | — |
| 21 |
| 15.92ms |
| — |
| hf-003 | 36.92ms | 0.141s | 44 | 15.12ms | — |
| hf-004 | 48.21ms | 0.123s | 42 | 15.92ms | — |
| hf-005 | 58.15ms | 0.104s | 20 | 15.55ms | — |
| hf-006 | 38.66ms | 0.157s | 46 | 16.67ms | — |
| hf-007 | 42.56ms | 0.161s | 43 | 15.10ms | — |
| hf-008 | 67.89ms | 0.254s | 51 | 48.36ms | — |
| hf-009 | 62.04ms | 0.151s | 28 | 15.09ms | — |
| hf-010 | 67.01ms | 0.161s | 43 | 17.57ms | — |
| hf-011 | 58.00ms | 0.194s | 52 | 15.48ms | — |
| hf-012 | 36.32ms | 0.080s | 23 | 15.25ms | — |
| hf-013 | 66.09ms | 0.173s | 43 | 15.94ms | — |
| hf-014 | 43.87ms | 0.134s | 48 | 15.14ms | — |
| hf-015 | 67.95ms | 0.250s | 42 | 44.26ms | — |
| hf-016 | 58.39ms | 0.148s | 43 | 16.65ms | — |
| hf-017 | 56.73ms | 0.177s | 43 | 15.32ms | — |
| hf-018 | 57.29ms | 0.163s | 44 | 15.80ms | — |
| hf-019 | 57.57ms | 0.194s | 54 | 15.65ms | — |
| hf-020 | 38.24ms | 0.142s | 41 | 15.12ms | — |
| hf-021 | 67.27ms | 0.161s | 47 | 18.53ms | — |
| hf-022 | 36.01ms | 0.125s | 48 | 15.57ms | — |
| hf-023 | 67.08ms | 0.192s | 52 | 17.79ms | — |
| hf-024 | 47.15ms | 0.166s | 46 | 15.30ms | — |
| hf-025 | 38.51ms | 0.127s | 42 | 15.11ms | — |
| hf-026 | 66.89ms | 0.158s | 23 | 15.59ms | — |
| hf-027 | 67.42ms | 0.162s | 38 | 20.14ms | — |
| hf-028 | 70.52ms | 0.194s | 46 | 17.46ms | — |
| hf-029 | 68.74ms | 0.163s | 46 | 18.64ms | — |
| hf-030 | 59.64ms | 0.151s | 41 | 15.38ms | — |
| hf-031 | 41.37ms | 0.117s | 42 | 15.27ms | — |
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