Round score
0.7174
+71.74%
Prompts scored
29 / 32
median_e2e_speedup
Below tolerance
3
answered too little to score
Zeroed total
3
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 | +71.35% | 0.733s | 0.210s | 42 | — |
| hf-001 | +73.41% | 0.684s | 0.182s | 40 | — |
| hf-002 | +78.17% | 0.687s | 0.150s |
32 requests
| Request | TTFT | Total | Tokens | Slowest gap | Stream |
|---|---|---|---|---|---|
| hf-000 | 97.21ms | 0.210s | 42 | 17.75ms | — |
| hf-001 | 101.24ms | 0.197s | 47 | 17.66ms | — |
| hf-002 | 70.35ms | 0.166s |
verdict
evidence
min logprob
num prompts
mean logprob
coverage ratio
quantile logprob
num positions scored
role
| 43 |
| — |
| hf-003 | +70.53% | 0.583s | 0.172s | 40 | — |
| hf-004 | +73.76% | 0.738s | 0.194s | 52 | — |
| hf-005 | +58.08% | 0.704s | 0.295s | 45 | — |
| hf-006 | +73.25% | 0.710s | 0.190s | 42 | — |
| hf-007 | +73.78% | 0.702s | 0.184s | 41 | — |
| hf-008 | +72.80% | 0.708s | 0.193s | 40 | — |
| hf-009 | +79.71% | 0.908s | 0.184s | 52 | — |
| hf-010 | — | — | — | 21 | candidate output below tolerance |
| hf-011 | +70.00% | 0.719s | 0.216s | 42 | — |
| hf-012 | +74.05% | 0.793s | 0.206s | 46 | — |
| hf-013 | +75.06% | 0.687s | 0.171s | 40 | — |
| hf-014 | +75.97% | 0.706s | 0.170s | 42 | — |
| hf-015 | +57.73% | 0.765s | 0.324s | 47 | — |
| hf-016 | +73.21% | 0.760s | 0.204s | 47 | — |
| hf-017 | +75.85% | 0.727s | 0.176s | 41 | — |
| hf-018 | +78.26% | 0.755s | 0.164s | 43 | — |
| hf-019 | +74.73% | 0.752s | 0.190s | 43 | — |
| hf-020 | +70.05% | 0.731s | 0.219s | 40 | — |
| hf-021 | +75.21% | 0.749s | 0.186s | 42 | — |
| hf-022 | +70.29% | 0.711s | 0.211s | 42 | — |
| hf-023 | — | — | — | 41 | candidate output below tolerance |
| hf-025 | +64.31% | 0.733s | 0.262s | 43 | — |
| hf-026 | +61.33% | 0.755s | 0.292s | 43 | — |
| hf-027 | +62.35% | 0.788s | 0.297s | 51 | — |
| hf-028 | +69.16% | 0.607s | 0.187s | 41 | — |
| hf-030 | +63.36% | 0.356s | 0.130s | 21 | — |
| hf-029 | +72.13% | 0.676s | 0.188s | 48 | — |
| hf-024 | +67.24% | 2.119s | 0.694s | 154 | — |
| hf-031 | — | — | — | 48 | candidate output below tolerance |
| 47 |
| 17.18ms |
| — |
| hf-003 | 75.66ms | 0.187s | 42 | 17.13ms | — |
| hf-004 | 65.80ms | 0.209s | 58 | 18.28ms | — |
| hf-005 | 80.53ms | 0.295s | 45 | 44.91ms | — |
| hf-006 | 95.37ms | 0.190s | 42 | 16.64ms | — |
| hf-007 | 103.04ms | 0.184s | 41 | 17.64ms | — |
| hf-008 | 65.92ms | 0.193s | 40 | 17.23ms | — |
| hf-009 | 57.26ms | 0.255s | 68 | 40.24ms | — |
| hf-010 | 90.05ms | 0.139s | 21 | 17.90ms | — |
| hf-011 | 103.03ms | 0.257s | 47 | 41.26ms | — |
| hf-012 | 94.59ms | 0.206s | 46 | 17.04ms | — |
| hf-013 | 75.81ms | 0.171s | 40 | 17.78ms | — |
| hf-014 | 74.42ms | 0.170s | 42 | 16.97ms | — |
| hf-015 | 131.41ms | 0.340s | 62 | 40.27ms | — |
| hf-016 | 90.16ms | 0.204s | 47 | 17.00ms | — |
| hf-017 | 79.85ms | 0.176s | 43 | 17.59ms | — |
| hf-018 | 68.72ms | 0.164s | 43 | 17.30ms | — |
| hf-019 | 92.08ms | 0.190s | 43 | 17.95ms | — |
| hf-020 | 121.90ms | 0.219s | 40 | 17.72ms | — |
| hf-021 | 89.91ms | 0.186s | 42 | 17.44ms | — |
| hf-022 | 83.08ms | 0.211s | 42 | 17.67ms | — |
| hf-023 | 106.72ms | 0.247s | 41 | 42.10ms | — |
| hf-024 | 81.88ms | 0.694s | 154 | 54.44ms | — |
| hf-025 | 87.19ms | 0.262s | 43 | 42.50ms | — |
| hf-026 | 113.77ms | 0.292s | 46 | 40.22ms | — |
| hf-027 | 91.73ms | 0.330s | 68 | 40.52ms | — |
| hf-028 | 90.63ms | 0.203s | 42 | 17.01ms | — |
| hf-029 | 76.59ms | 0.246s | 67 | 41.96ms | — |
| hf-030 | 81.82ms | 0.177s | 47 | 17.65ms | — |
| hf-031 | 63.27ms | 0.221s | 48 | 17.42ms | — |
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