Round score
—
degenerate output (baseline-relative repetition failures exceed allowance of 4 prompts (hf-008: distinct 16-gram ratio 0.670 is more than 0.10 below baseline 0.956 over 21310 chars))
Eligible requests
—
trusted baseline exclusions removed
Failed eligible requests
—
each request counted once
Raw weighted speedup
—
before failure cap and deduction
Temperature range
Request concurrency
Tier scheduling
Output policy
Thinking
Context limit
The same eligible inputs and output ceilings apply to every engine. Candidates must emit at least 90% of baseline tokens per request; natural output lengths can affect tier completion time. Slots refill until the group drains and stay occupied through protocol completion. Baseline exclusions and final drain can reduce actual concurrency below the configured cap.
The round score uses weighted completion times for all eligible work in each tier. Per-request speedups below are diagnostics, not score contributions. Only trusted baseline exclusions are removed; candidate failures remain in the eligible count and incur the configured failure penalty.
No prompt scores
degenerate output (baseline-relative repetition failures exceed allowance of 4 prompts (hf-008: distinct 16-gram ratio 0.670 is more than 0.10 below baseline 0.956 over 21310 chars))
16 requests
| Request | TTFT | Total | Tokens | Slowest gap | Stream |
|---|---|---|---|---|---|
| hf-000 | 158.32ms | 14.802s | 3250 | 18.17ms | 1010 gaps for 3250 tokens |
| hf-001 | 146.63ms | 22.232s | 4467 | 45.55ms | 1515 gaps for 4467 tokens |
| hf-002 | 119.12ms | 18.674s |
Occupied client slots include protocol completion and final drain. Mean occupancy is weighted by elapsed time; it can be below the requested cap.
| Repetition | Group start | Requested | Effective cap | Peak | Mean |
|---|---|---|---|---|---|
| 1 | 0.000s | C4 | 4 | 4 | 3.47 |
| 1 | 48.396s | C4 | 4 | 4 | 3.28 |
| 2 | 0.000s | C4 | 4 | 4 | 3.72 |
| 2 | 46.645s | C4 | 4 | 4 | 3.51 |
| 3 | 0.000s | C4 | 4 | 4 | 3.88 |
| 3 | 45.687s | C4 | 4 | 4 | 3.57 |
| 3851 |
| 46.39ms |
| 1328 gaps for 3851 tokens |
| hf-003 | 126.35ms | 25.916s | 5120 | 50.40ms | 1760 gaps for 5120 tokens |
| hf-004 | 161.47ms | 15.970s | 3306 | 44.29ms | 1088 gaps for 3306 tokens |
| hf-005 | 101.82ms | 18.410s | 3914 | 48.49ms | 1244 gaps for 3914 tokens |
| hf-006 | 91.11ms | 25.532s | 5120 | 43.90ms | 1783 gaps for 5120 tokens |
| hf-007 | 99.89ms | 23.415s | 5120 | 44.62ms | 1663 gaps for 5120 tokens |
| hf-008 | 88.12ms | 23.460s | 5120 | 46.02ms | 1646 gaps for 5120 tokens |
| hf-009 | 112.93ms | 21.463s | 4826 | 47.32ms | 1520 gaps for 4826 tokens |
| hf-010 | 98.88ms | 21.436s | 4616 | 44.59ms | 1525 gaps for 4616 tokens |
| hf-011 | 84.38ms | 22.246s | 5120 | 16.40ms | 1572 gaps for 5120 tokens |
| hf-012 | 74.84ms | 21.985s | 4665 | 44.82ms | 1557 gaps for 4665 tokens |
| hf-013 | 73.48ms | 22.760s | 5120 | 42.15ms | 1617 gaps for 5120 tokens |
| hf-014 | 71.38ms | 23.368s | 4986 | 42.01ms | 1677 gaps for 4986 tokens |
| hf-015 | 76.92ms | 23.037s | 5120 | 16.40ms | 1664 gaps for 5120 tokens |
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 | 0.7621 | 0.9835 | 0.2215 | 0.1000 | Failed distinct 16-gram ratio 0.762 is more than 0.10 below baseline 0.984 over 14464 chars |
| hf-001 | 0.9209 | 0.9535 | 0.0326 | 0.1000 | Passed |
| hf-002 | 0.8921 | 0.8763 | -0.0158 | 0.1000 | Passed |
| hf-003 | 0.7335 | 0.8895 | 0.1560 | 0.1000 | Failed distinct 16-gram ratio 0.733 is more than 0.10 below baseline 0.889 over 22159 chars |
| hf-004 | 0.9452 | 0.9121 | -0.0331 | 0.1000 | Passed |
| hf-005 | 0.7797 | 0.9532 | 0.1735 | 0.1000 | Failed rep 2: longest repeated span ratio 0.292 at or above harness ceiling over 15465 chars |
| hf-006 | 0.9316 | 0.9494 | 0.0178 | 0.1000 | Passed |
| hf-007 | 0.6668 | 0.9088 | 0.2420 | 0.1000 | Failed longest repeated span ratio 0.256 at or above harness ceiling over 21226 chars |
| hf-008 | 0.6698 | 0.9557 | 0.2859 | 0.1000 | Failed rep 3: longest repeated span ratio 0.272 at or above harness ceiling over 20766 chars |
| hf-009 | 0.8004 | 0.8497 | 0.0492 | 0.1000 | Passed |
| hf-010 | 0.9291 | 0.9712 | 0.0421 | 0.1000 | Passed |
| hf-011 | 0.5911 | 0.8315 | 0.2404 | 0.1000 | Failed distinct 16-gram ratio 0.591 is more than 0.10 below baseline 0.832 over 21651 chars |
| hf-012 | 0.8718 | 0.9198 | 0.0480 | 0.1000 | Failed rep 2: longest repeated span ratio 0.399 at or above harness ceiling over 22271 chars |
| hf-013 | 0.7390 | 0.9577 | 0.2187 | 0.1000 | Failed distinct 16-gram ratio 0.739 is more than 0.10 below baseline 0.958 over 22329 chars |
| hf-014 | 0.9245 | 0.9436 | 0.0191 | 0.1000 | Passed |
| hf-015 | 0.8193 | 0.9561 | 0.1367 | 0.1000 | Failed rep 3: longest repeated span ratio 0.498 at or above harness ceiling over 21072 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-010 | 1 | 0.855638 | 1 | 0 |
| hf-002 | 1 | 0.189667 | 1 | 0 |
| hf-009 | 1 | 0.208989 | 1 | 0 |
| hf-007 | 1 | 0.417374 | 1 | 0 |
| hf-012 | 1 | 0.325767 | 1 | 0 |
| hf-013 | 1 | 0.566642 | 1 | 0 |
| hf-014 | 1 | 0.590278 | 1 | 0 |
| hf-015 | 1 | 0.591114 | 1 | 0 |
| hf-000 | 1 | 0.91753 | 1 | 0 |
| hf-004 | 1 | 0.84741 | 1 | 0 |
| hf-001 | 1 | 0.486939 | 1 | 0 |
| hf-003 | 1 | 0.159301 | 1 | 0 |
| hf-006 | 1 | 0.418659 | 1 | 0 |
| hf-005 | 1 | 0.743609 | 1 | 0 |
| hf-008 | 1 | 0.355751 | 1 | 0 |
| hf-011 | 1 | 0.433119 | 1 | 0 |
| hf-002 | 2 | 0.189667 | 1 | 0 |
| hf-009 | 2 | 0.208989 | 1 | 0 |
| hf-010 | 2 | 0.855638 | 1 | 0 |
| hf-007 | 2 | 0.417374 | 1 | 0 |
| hf-012 | 2 | 0.325767 | 1 | 0 |
| hf-013 | 2 | 0.566642 | 1 | 0 |
| hf-014 | 2 | 0.590278 | 1 | 0 |
| hf-015 | 2 | 0.591114 | 1 | 0 |
| hf-000 | 2 | 0.91753 | 1 | 0 |
| hf-004 | 2 | 0.84741 | 1 | 0 |
| hf-001 | 2 | 0.486939 | 1 | 0 |
| hf-003 | 2 | 0.159301 | 1 | 0 |
| hf-005 | 2 | 0.743609 | 1 | 0 |
| hf-006 | 2 | 0.418659 | 1 | 0 |
| hf-008 | 2 | 0.355751 | 1 | 0 |
| hf-011 | 2 | 0.433119 | 1 | 0 |
| hf-002 | 3 | 0.189667 | 1 | 0 |
| hf-007 | 3 | 0.417374 | 1 | 0 |
| hf-009 | 3 | 0.208989 | 1 | 0 |
| hf-010 | 3 | 0.855638 | 1 | 0 |
| hf-012 | 3 | 0.325767 | 1 | 0 |
| hf-013 | 3 | 0.566642 | 1 | 0 |
| hf-015 | 3 | 0.591114 | 1 | 0 |
| hf-014 | 3 | 0.590278 | 1 | 0 |
| hf-004 | 3 | 0.84741 | 1 | 0 |
| hf-000 | 3 | 0.91753 | 1 | 0 |
| hf-001 | 3 | 0.486939 | 1 | 0 |
| hf-003 | 3 | 0.159301 | 1 | 0 |
| hf-005 | 3 | 0.743609 | 1 | 0 |
| hf-006 | 3 | 0.418659 | 1 | 0 |
| hf-011 | 3 | 0.433119 | 1 | 0 |
| hf-008 | 3 | 0.355751 | 1 | 0 |
reason
verdict
evidence
min logprob
num prompts
mean logprob
coverage ratio
quantile logprob
num positions scored
role
evidence
request concurrency
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
tier completion max rel range
Image digest
Patch hash