Round score
—
degenerate output (hf-002: rep 1: distinct 16-gram ratio 0.065 below harness floor over 22412 chars)
Prompts scored
0 / 0
median_e2e_speedup
Below tolerance
0
answered too little to score
Zeroed total
0
scored 0.0 without measuring
Temperature range
Request interval
Thinking
Context limit
The same inputs and output limits apply to every engine. Actual concurrency depends on response duration and engine scheduling.
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.
No prompt scores
degenerate output (hf-002: rep 1: distinct 16-gram ratio 0.065 below harness floor over 22412 chars)
32 requests
| Request | TTFT | Total | Tokens | Slowest gap | Stream |
|---|---|---|---|---|---|
| hf-000 | 363.72ms | 31.538s | 4637 | 156.00ms | — |
| hf-001 | 343.75ms | 70.756s | 4630 | 201.41ms | — |
| hf-002 | 360.49ms | 32.691s |
Each latency-median response is compared with the lowest valid opening-baseline ratio for the same prompt. Drops above the recorded limit fail; absolute checks also apply. These checks detect repetition, not overall writing quality.
| Prompt | Candidate char-16 | Baseline char-16 | Drop | Limit | Result |
|---|---|---|---|---|---|
| hf-000 | Not recorded | Not recorded | Not recorded | Not recorded | Excluded baseline-drift: longest repeated span ratio 0.358 at or above harness ceiling over 20844 chars |
| hf-001 | 0.7981 | 0.9534 | 0.1553 | 0.1000 | Failed distinct 16-gram ratio 0.798 is more than 0.10 below baseline 0.953 over 7257 chars |
| hf-002 | 0.1873 | 0.9356 | 0.7483 | 0.1000 |
| 4687 |
| 377.58ms |
| — |
| hf-003 | 445.72ms | 31.015s | 4620 | 163.39ms | — |
| hf-004 | 770.96ms | 34.357s | 4615 | 177.10ms | — |
| hf-005 | 441.78ms | 34.107s | 4633 | 171.05ms | — |
| hf-006 | 446.12ms | 63.819s | 4765 | 174.87ms | — |
| hf-007 | 557.21ms | 93.106s | 5120 | 169.40ms | — |
| hf-008 | 763.38ms | 32.440s | 4701 | 173.94ms | — |
| hf-009 | 553.23ms | 31.242s | 4631 | 168.65ms | — |
| hf-010 | 760.05ms | 32.849s | 4809 | 173.70ms | — |
| hf-011 | 429.68ms | 38.894s | 5120 | 170.92ms | — |
| hf-012 | 434.54ms | 34.570s | 4614 | 174.82ms | — |
| hf-013 | 425.78ms | 31.376s | 4634 | 170.58ms | — |
| hf-014 | 424.70ms | 32.901s | 4629 | 172.23ms | — |
| hf-015 | 1.352s | 36.553s | 4637 | 170.83ms | — |
| hf-016 | 1.367s | 31.675s | 4615 | 42.60ms | — |
| hf-017 | 1.348s | 33.800s | 4618 | 170.81ms | — |
| hf-018 | 1.369s | 32.125s | 4714 | 174.90ms | — |
| hf-019 | 1.356s | 71.965s | 4779 | 622.40ms | — |
| hf-020 | 1.365s | 34.948s | 4632 | 178.66ms | — |
| hf-021 | 1.351s | 32.166s | 4645 | 622.78ms | — |
| hf-022 | 1.355s | 32.646s | 4661 | 172.43ms | — |
| hf-023 | 1.494s | 32.073s | 4622 | 169.17ms | — |
| hf-024 | 32.341s | 59.614s | 4879 | 31.05ms | — |
| hf-025 | 31.984s | 120.027s | 4612 | 167.67ms | — |
| hf-026 | 31.992s | 66.153s | 5120 | 160.81ms | — |
| hf-027 | 33.191s | 62.487s | 4621 | 167.56ms | — |
| hf-028 | 31.988s | 77.982s | 4686 | 160.40ms | — |
| hf-029 | 33.187s | 62.664s | 4659 | 167.63ms | — |
| hf-030 | 32.329s | 59.603s | 4706 | 30.73ms | — |
| hf-031 | 32.364s | 64.208s | 4612 | 35.75ms | — |
rep 1: distinct 16-gram ratio 0.065 below harness floor over 22412 chars
| hf-003 | 0.0722 | 0.7501 | 0.6779 | 0.1000 | Failed distinct 16-gram ratio 0.072 below harness floor over 6562 chars |
| hf-004 | 0.0584 | 0.9438 | 0.8854 | 0.1000 | Failed distinct 16-gram ratio 0.058 below harness floor over 21488 chars |
| hf-005 | 0.0710 | 0.8907 | 0.8196 | 0.1000 | Failed distinct 16-gram ratio 0.075 below harness floor over 17365 chars |
| hf-006 | 0.8371 | 0.9680 | 0.1309 | 0.1000 | Failed distinct 16-gram ratio 0.837 is more than 0.10 below baseline 0.968 over 10889 chars |
| hf-007 | 0.6021 | 0.9748 | 0.3727 | 0.1000 | Failed distinct 16-gram ratio 0.602 is more than 0.10 below baseline 0.975 over 16358 chars |
| hf-008 | Not recorded | Not recorded | Not recorded | Not recorded | Excluded baseline: longest repeated span ratio 0.493 at or above harness ceiling over 21096 chars |
| hf-009 | 0.0843 | 0.8367 | 0.7524 | 0.1000 | Failed distinct 16-gram ratio 0.084 below harness floor over 16187 chars |
| hf-010 | 0.0958 | 0.8625 | 0.7667 | 0.1000 | Failed distinct 16-gram ratio 0.099 below harness floor over 8762 chars |
| hf-011 | 0.3524 | 0.9686 | 0.6162 | 0.1000 | Failed distinct 16-gram ratio 0.352 is more than 0.10 below baseline 0.969 over 5841 chars |
| hf-012 | 0.0466 | 0.9148 | 0.8682 | 0.1000 | Failed distinct 16-gram ratio 0.047 below harness floor over 12250 chars |
| hf-013 | 0.1308 | 0.9192 | 0.7884 | 0.1000 | Failed distinct 16-gram ratio 0.131 below harness floor over 9251 chars |
| hf-014 | 0.2278 | 0.9799 | 0.7521 | 0.1000 | Failed rep 1: distinct 16-gram ratio 0.084 below harness floor over 11722 chars |
| hf-015 | 0.0836 | 0.9597 | 0.8760 | 0.1000 | Failed distinct 16-gram ratio 0.084 below harness floor over 10515 chars |
| hf-016 | 0.0428 | 0.9300 | 0.8872 | 0.1000 | Failed distinct 16-gram ratio 0.043 below harness floor over 17846 chars |
| hf-017 | 0.0654 | 0.9093 | 0.8439 | 0.1000 | Failed distinct 16-gram ratio 0.063 below harness floor over 10925 chars |
| hf-018 | 0.0575 | 0.6961 | 0.6386 | 0.1000 | Failed distinct 16-gram ratio 0.059 below harness floor over 16244 chars |
| hf-019 | 0.8393 | 0.9697 | 0.1303 | 0.1000 | Failed distinct 16-gram ratio 0.839 is more than 0.10 below baseline 0.970 over 9114 chars |
| hf-020 | 0.0588 | 0.9413 | 0.8825 | 0.1000 | Failed distinct 16-gram ratio 0.059 below harness floor over 14887 chars |
| hf-021 | 0.0473 | 0.8865 | 0.8392 | 0.1000 | Failed distinct 16-gram ratio 0.047 below harness floor over 16921 chars |
| hf-022 | 0.0677 | 0.9377 | 0.8700 | 0.1000 | Failed distinct 16-gram ratio 0.068 below harness floor over 10896 chars |
| hf-023 | 0.0920 | 0.9276 | 0.8357 | 0.1000 | Failed distinct 16-gram ratio 0.092 below harness floor over 9408 chars |
| hf-024 | 0.0803 | 0.8967 | 0.8164 | 0.1000 | Failed distinct 16-gram ratio 0.082 below harness floor over 9312 chars |
| hf-025 | 0.8738 | 0.9667 | 0.0928 | 0.1000 | Passed |
| hf-026 | 0.2629 | 0.9004 | 0.6375 | 0.1000 | Failed distinct 16-gram ratio 0.263 is more than 0.10 below baseline 0.900 over 22024 chars |
| hf-027 | 0.1287 | 0.9665 | 0.8378 | 0.1000 | Failed distinct 16-gram ratio 0.129 below harness floor over 9588 chars |
| hf-028 | 0.7750 | 0.9810 | 0.2060 | 0.1000 | Failed distinct 16-gram ratio 0.775 is more than 0.10 below baseline 0.981 over 8571 chars |
| hf-029 | 0.2069 | 0.8986 | 0.6918 | 0.1000 | Failed rep 1: distinct 16-gram ratio 0.034 below harness floor over 21271 chars |
| hf-030 | 0.0686 | 0.9426 | 0.8740 | 0.1000 | Failed distinct 16-gram ratio 0.070 below harness floor over 18489 chars |
| hf-031 | 0.0833 | 0.9663 | 0.8830 | 0.1000 | Failed distinct 16-gram ratio 0.083 below harness floor over 20827 chars |
Actual settings for each measured request. Each prompt keeps its temperature across repetitions and engines. Generation uses seed 0 throughout qualification, warmups and measured repetitions; fixed settings do not guarantee identical output text.
| Prompt | Repetition | Temperature | Top p | Seed |
|---|---|---|---|---|
| hf-008 | 1 | 0.210727 | 1 | 0 |
| hf-021 | 1 | 0.322792 | 1 | 0 |
| hf-009 | 1 | 0.629894 | 1 | 0 |
| hf-003 | 1 | 0.22992 | 1 | 0 |
| hf-013 | 1 | 0.642291 | 1 | 0 |
| hf-014 | 1 | 0.830809 | 1 | 0 |
| hf-023 | 1 | 0.394813 | 1 | 0 |
| hf-016 | 1 | 0.193595 | 1 | 0 |
| hf-022 | 1 | 0.599649 | 1 | 0 |
| hf-018 | 1 | 0.267796 | 1 | 0 |
| hf-000 | 1 | 0.151884 | 1 | 0 |
| hf-002 | 1 | 0.672959 | 1 | 0 |
| hf-017 | 1 | 0.188127 | 1 | 0 |
| hf-004 | 1 | 0.232513 | 1 | 0 |
| hf-005 | 1 | 0.313157 | 1 | 0 |
| hf-012 | 1 | 0.23919 | 1 | 0 |
| hf-020 | 1 | 0.471604 | 1 | 0 |
| hf-015 | 1 | 0.778185 | 1 | 0 |
| hf-010 | 1 | 0.731621 | 1 | 0 |
| hf-011 | 1 | 0.87649 | 1 | 0 |
| hf-024 | 1 | 0.590361 | 1 | 0 |
| hf-030 | 1 | 0.354911 | 1 | 0 |
| hf-029 | 1 | 0.491639 | 1 | 0 |
| hf-006 | 1 | 1.002486 | 1 | 0 |
| hf-031 | 1 | 0.695718 | 1 | 0 |
| hf-026 | 1 | 0.550558 | 1 | 0 |
| hf-027 | 1 | 0.690363 | 1 | 0 |
| hf-001 | 1 | 0.971955 | 1 | 0 |
| hf-007 | 1 | 0.998372 | 1 | 0 |
| hf-028 | 1 | 0.961174 | 1 | 0 |
| hf-019 | 1 | 0.963234 | 1 | 0 |
| hf-025 | 1 | 1.004332 | 1 | 0 |
| hf-003 | 2 | 0.22992 | 1 | 0 |
| hf-018 | 2 | 0.267796 | 1 | 0 |
| hf-012 | 2 | 0.23919 | 1 | 0 |
| hf-000 | 2 | 0.151884 | 1 | 0 |
| hf-016 | 2 | 0.193595 | 1 | 0 |
| hf-002 | 2 | 0.672959 | 1 | 0 |
| hf-009 | 2 | 0.629894 | 1 | 0 |
| hf-021 | 2 | 0.322792 | 1 | 0 |
| hf-020 | 2 | 0.471604 | 1 | 0 |
| hf-010 | 2 | 0.731621 | 1 | 0 |
| hf-022 | 2 | 0.599649 | 1 | 0 |
| hf-015 | 2 | 0.778185 | 1 | 0 |
| hf-014 | 2 | 0.830809 | 1 | 0 |
| hf-013 | 2 | 0.642291 | 1 | 0 |
| hf-008 | 2 | 0.210727 | 1 | 0 |
| hf-005 | 2 | 0.313157 | 1 | 0 |
| hf-023 | 2 | 0.394813 | 1 | 0 |
| hf-011 | 2 | 0.87649 | 1 | 0 |
| hf-017 | 2 | 0.188127 | 1 | 0 |
| hf-004 | 2 | 0.232513 | 1 | 0 |
| hf-030 | 2 | 0.354911 | 1 | 0 |
| hf-024 | 2 | 0.590361 | 1 | 0 |
| hf-027 | 2 | 0.690363 | 1 | 0 |
| hf-029 | 2 | 0.491639 | 1 | 0 |
| hf-031 | 2 | 0.695718 | 1 | 0 |
| hf-006 | 2 | 1.002486 | 1 | 0 |
| hf-001 | 2 | 0.971955 | 1 | 0 |
| hf-026 | 2 | 0.550558 | 1 | 0 |
| hf-019 | 2 | 0.963234 | 1 | 0 |
| hf-028 | 2 | 0.961174 | 1 | 0 |
| hf-007 | 2 | 0.998372 | 1 | 0 |
| hf-025 | 2 | 1.004332 | 1 | 0 |
| hf-013 | 3 | 0.642291 | 1 | 0 |
| hf-005 | 3 | 0.313157 | 1 | 0 |
| hf-009 | 3 | 0.629894 | 1 | 0 |
| hf-000 | 3 | 0.151884 | 1 | 0 |
| hf-017 | 3 | 0.188127 | 1 | 0 |
| hf-003 | 3 | 0.22992 | 1 | 0 |
| hf-023 | 3 | 0.394813 | 1 | 0 |
| hf-008 | 3 | 0.210727 | 1 | 0 |
| hf-002 | 3 | 0.672959 | 1 | 0 |
| hf-010 | 3 | 0.731621 | 1 | 0 |
| hf-018 | 3 | 0.267796 | 1 | 0 |
| hf-016 | 3 | 0.193595 | 1 | 0 |
| hf-022 | 3 | 0.599649 | 1 | 0 |
| hf-004 | 3 | 0.232513 | 1 | 0 |
| hf-014 | 3 | 0.830809 | 1 | 0 |
| hf-011 | 3 | 0.87649 | 1 | 0 |
| hf-012 | 3 | 0.23919 | 1 | 0 |
| hf-015 | 3 | 0.778185 | 1 | 0 |
| hf-019 | 3 | 0.963234 | 1 | 0 |
| hf-021 | 3 | 0.322792 | 1 | 0 |
| hf-020 | 3 | 0.471604 | 1 | 0 |
| hf-027 | 3 | 0.690363 | 1 | 0 |
| hf-006 | 3 | 1.002486 | 1 | 0 |
| hf-026 | 3 | 0.550558 | 1 | 0 |
| hf-030 | 3 | 0.354911 | 1 | 0 |
| hf-029 | 3 | 0.491639 | 1 | 0 |
| hf-031 | 3 | 0.695718 | 1 | 0 |
| hf-024 | 3 | 0.590361 | 1 | 0 |
| hf-001 | 3 | 0.971955 | 1 | 0 |
| hf-007 | 3 | 0.998372 | 1 | 0 |
| hf-028 | 3 | 0.961174 | 1 | 0 |
| hf-025 | 3 | 1.004332 | 1 | 0 |
reason
verdict
evidence
min logprob
num prompts
mean logprob
coverage ratio
quantile logprob
num positions scored
role
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