Round score
0.2577
weighted tier speedup after failure cap and deduction
Eligible requests
27
trusted baseline exclusions removed
Failed eligible requests
10
each request counted once
Raw weighted speedup
+29.47%
before failure cap and deduction
Raw weighted speedup
Eligible speedup
Request failures cap speed credit at zero: min(0.294721, 0) = 0.294721. Negative speedups remain negative.
Failed requests
Deduction
Final score
Each duration is the median across repetitions, from the group start until the tier's last request finishes. This includes client queueing. Weighted speedup = sum of weight × (1 − candidate / baseline).
| Input tier | Eligible requests | Weight | Baseline | Candidate | Speedup | Contribution |
|---|---|---|---|---|---|---|
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.
27 prompts
| Prompt | Diagnostic speedup | Baseline | Candidate | Aligned output | Diagnostic / failure |
|---|---|---|---|---|---|
| hf-004 | — | — | — | 2404 | candidate output below tolerance |
| hf-001 | — | — | — | 4538 | Per-token timing unavailable (not a request failure) |
| hf-015 | — | — | — |
27 requests
| Request | TTFT | Total | Tokens | Slowest gap | Stream |
|---|---|---|---|---|---|
| hf-000 | 1.908s | 49.691s | 3108 | 34.24ms | 1669 gaps for 3108 tokens |
| hf-001 | 1.907s | 62.184s | 5120 | 35.07ms | 2158 gaps for 5120 tokens |
| hf-002 | 1.843s | 72.897s |
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 | C32 | 27 | 27 | 20.99 |
| 2 | 0.000s | C32 | 27 | 27 | 22.47 |
| 3 | 0.000s | C32 | 27 | 27 | 21.91 |
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 |
|---|
| 16k |
|---|
| 6 |
| 25% |
| 105.756s |
| 73.771s |
| +30.24% |
| +7.56% |
| 32k | 7 | 25% | 114.915s | 81.676s | +28.92% | +7.23% |
|---|
| 64k | 8 | 25% | 113.059s | 86.175s | +23.78% | +5.94% |
|---|
| 128k | 6 | 25% | 114.480s | 74.481s | +34.94% | +8.73% |
|---|
| 2761 |
| Per-token timing unavailable (not a request failure) |
| hf-028 | — | — | — | 4078 | Per-token timing unavailable (not a request failure) |
| hf-010 | — | — | — | 3013 | Per-token timing unavailable (not a request failure) |
| hf-018 | — | — | — | 3790 | Per-token timing unavailable (not a request failure) |
| hf-017 | — | — | — | 3665 | candidate output below tolerance |
| hf-026 | — | — | — | 3948 | Per-token timing unavailable (not a request failure) |
| hf-019 | — | — | — | 3842 | candidate output below tolerance |
| hf-029 | — | — | — | 4284 | candidate output below tolerance |
| hf-013 | — | — | — | 5120 | Per-token timing unavailable (not a request failure) |
| hf-014 | — | — | — | 5120 | Per-token timing unavailable (not a request failure) |
| hf-021 | — | — | — | 4288 | Per-token timing unavailable (not a request failure) |
| hf-009 | — | — | — | 2944 | candidate output below tolerance |
| hf-011 | — | — | — | 3237 | Per-token timing unavailable (not a request failure) |
| hf-000 | — | — | — | 3056 | Per-token timing unavailable (not a request failure) |
| hf-007 | — | — | — | 4196 | Per-token timing unavailable (not a request failure) |
| hf-008 | — | — | — | 3572 | candidate output below tolerance |
| hf-031 | — | — | — | 1737 | candidate output below tolerance |
| hf-025 | — | — | — | 3078 | candidate output below tolerance |
| hf-006 | — | — | — | 4933 | Per-token timing unavailable (not a request failure) |
| hf-003 | — | — | — | 5120 | Per-token timing unavailable (not a request failure) |
| hf-030 | — | — | — | 2698 | candidate output below tolerance |
| hf-002 | — | — | — | 5120 | Per-token timing unavailable (not a request failure) |
| hf-024 | — | — | — | 4063 | Per-token timing unavailable (not a request failure) |
| hf-027 | — | — | — | 5120 | Per-token timing unavailable (not a request failure) |
| hf-012 | — | — | — | 3669 | candidate output below tolerance |
| 5120 |
| 32.54ms |
| 2854 gaps for 5120 tokens |
| hf-003 | 1.831s | 67.041s | 5120 | 32.52ms | 2420 gaps for 5120 tokens |
| hf-004 | 1.837s | 38.905s | 2404 | 32.59ms | 1293 gaps for 2404 tokens |
| hf-006 | 1.836s | 68.796s | 4933 | 33.55ms | 2525 gaps for 4933 tokens |
| hf-007 | 1.905s | 68.599s | 4338 | 35.18ms | 2460 gaps for 4338 tokens |
| hf-008 | 1.906s | 53.097s | 3572 | 34.97ms | 1795 gaps for 3572 tokens |
| hf-009 | 1.841s | 50.032s | 2944 | 32.55ms | 1696 gaps for 2944 tokens |
| hf-010 | 1.839s | 63.338s | 4731 | 33.48ms | 2236 gaps for 4731 tokens |
| hf-011 | 1.900s | 59.445s | 3237 | 35.15ms | 2044 gaps for 3237 tokens |
| hf-012 | 1.831s | 62.522s | 3669 | 33.41ms | 2200 gaps for 3669 tokens |
| hf-013 | 1.834s | 65.589s | 5120 | 33.18ms | 2343 gaps for 5120 tokens |
| hf-014 | 1.898s | 66.696s | 5120 | 35.08ms | 2361 gaps for 5120 tokens |
| hf-015 | 1.827s | 56.491s | 3361 | 33.12ms | 1947 gaps for 3361 tokens |
| hf-017 | 1.835s | 48.711s | 3665 | 33.38ms | 1647 gaps for 3665 tokens |
| hf-018 | 1.847s | 61.715s | 5120 | 40.31ms | 2245 gaps for 5120 tokens |
| hf-019 | 1.819s | 65.094s | 3842 | 33.43ms | 2318 gaps for 3842 tokens |
| hf-021 | 1.851s | 64.411s | 4288 | 40.31ms | 2392 gaps for 4288 tokens |
| hf-024 | 1.887s | 72.105s | 4751 | 35.04ms | 2666 gaps for 4751 tokens |
| hf-025 | 1.856s | 53.311s | 3078 | 40.37ms | 1880 gaps for 3078 tokens |
| hf-026 | 1.829s | 69.622s | 5120 | 33.23ms | 2580 gaps for 5120 tokens |
| hf-027 | 1.821s | 74.450s | 5120 | 33.44ms | 3020 gaps for 5120 tokens |
| hf-028 | 1.824s | 69.926s | 5120 | 33.11ms | 2603 gaps for 5120 tokens |
| hf-029 | 1.852s | 55.566s | 4284 | 40.30ms | 1973 gaps for 4284 tokens |
| hf-030 | 1.822s | 45.027s | 2698 | 33.12ms | 1512 gaps for 2698 tokens |
| hf-031 | 1.846s | 25.436s | 1737 | 40.31ms | 843 gaps for 1737 tokens |
| Baseline char-16 |
|---|
| Drop |
|---|
| Limit |
|---|
| Result |
|---|
| hf-000 | 0.9777 | 0.8609 | -0.1168 | 0.1000 | Passed |
| hf-001 | 0.7851 | 0.7593 | -0.0258 | 0.1000 | Passed |
| hf-002 | 0.7939 | 0.8267 | 0.0329 | 0.1000 | Passed |
| hf-003 | 0.6175 | 0.6962 | 0.0787 | 0.1000 | Passed |
| hf-004 | 0.8126 | 0.7489 | -0.0637 | 0.1000 | Passed |
| hf-006 | 0.7467 | 0.7215 | -0.0252 | 0.1000 | Passed |
| hf-007 | 0.7575 | 0.7442 | -0.0132 | 0.1000 | Passed |
| hf-008 | 0.8549 | 0.7113 | -0.1436 | 0.1000 | Passed |
| hf-009 | 0.7832 | 0.7206 | -0.0626 | 0.1000 | Passed |
| hf-010 | 0.9527 | 0.9460 | -0.0067 | 0.1000 | Passed |
| hf-011 | 0.9734 | 0.9592 | -0.0142 | 0.1000 | Passed |
| hf-012 | 0.8080 | 0.7730 | -0.0349 | 0.1000 | Passed |
| hf-013 | 0.8234 | 0.6721 | -0.1512 | 0.1000 | Passed |
| hf-014 | 0.7925 | 0.7972 | 0.0047 | 0.1000 | Passed |
| hf-015 | 0.7931 | 0.7775 | -0.0156 | 0.1000 | Passed |
| hf-017 | 0.7966 | 0.8522 | 0.0556 | 0.1000 | Passed |
| hf-018 | 0.7620 | 0.8040 | 0.0420 | 0.1000 | Passed |
| hf-019 | 0.9382 | 0.7187 | -0.2195 | 0.1000 | Passed |
| hf-021 | 0.9201 | 0.8080 | -0.1121 | 0.1000 | Passed |
| hf-024 | 0.7236 | 0.7532 | 0.0295 | 0.1000 | Passed |
| hf-025 | 0.8918 | 0.7077 | -0.1840 | 0.1000 | Passed |
| hf-026 | 0.7413 | 0.6506 | -0.0907 | 0.1000 | Passed |
| hf-027 | 0.8029 | 0.7536 | -0.0494 | 0.1000 | Passed |
| hf-028 | 0.8577 | 0.7202 | -0.1375 | 0.1000 | Passed |
| hf-029 | 0.9077 | 0.8660 | -0.0416 | 0.1000 | Passed |
| hf-030 | 0.9358 | 0.8274 | -0.1083 | 0.1000 | Passed |
| hf-031 | 0.9177 | 0.9619 | 0.0442 | 0.1000 | Passed |
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-000 | 1 | 1.005729 | 1 | 0 |
| hf-031 | 1 | 0.952171 | 1 | 0 |
| hf-017 | 1 | 0.880366 | 1 | 0 |
| hf-030 | 1 | 0.378947 | 1 | 0 |
| hf-008 | 1 | 0.276091 | 1 | 0 |
| hf-009 | 1 | 0.197127 | 1 | 0 |
| hf-015 | 1 | 0.197416 | 1 | 0 |
| hf-010 | 1 | 0.955022 | 1 | 0 |
| hf-025 | 1 | 0.139681 | 1 | 0 |
| hf-029 | 1 | 0.91998 | 1 | 0 |
| hf-012 | 1 | 0.116817 | 1 | 0 |
| hf-011 | 1 | 0.239254 | 1 | 0 |
| hf-028 | 1 | 0.506323 | 1 | 0 |
| hf-014 | 1 | 0.826373 | 1 | 0 |
| hf-018 | 1 | 0.925935 | 1 | 0 |
| hf-001 | 1 | 0.987002 | 1 | 0 |
| hf-013 | 1 | 0.755164 | 1 | 0 |
| hf-021 | 1 | 0.303212 | 1 | 0 |
| hf-003 | 1 | 0.24968 | 1 | 0 |
| hf-007 | 1 | 0.267166 | 1 | 0 |
| hf-026 | 1 | 0.6188 | 1 | 0 |
| hf-006 | 1 | 0.561965 | 1 | 0 |
| hf-004 | 1 | 0.374735 | 1 | 0 |
| hf-002 | 1 | 0.322905 | 1 | 0 |
| hf-019 | 1 | 0.26784 | 1 | 0 |
| hf-024 | 1 | 0.382038 | 1 | 0 |
| hf-027 | 1 | 0.15659 | 1 | 0 |
| hf-004 | 2 | 0.374735 | 1 | 0 |
| hf-031 | 2 | 0.952171 | 1 | 0 |
| hf-000 | 2 | 1.005729 | 1 | 0 |
| hf-008 | 2 | 0.276091 | 1 | 0 |
| hf-021 | 2 | 0.303212 | 1 | 0 |
| hf-009 | 2 | 0.197127 | 1 | 0 |
| hf-019 | 2 | 0.26784 | 1 | 0 |
| hf-025 | 2 | 0.139681 | 1 | 0 |
| hf-011 | 2 | 0.239254 | 1 | 0 |
| hf-029 | 2 | 0.91998 | 1 | 0 |
| hf-018 | 2 | 0.925935 | 1 | 0 |
| hf-001 | 2 | 0.987002 | 1 | 0 |
| hf-017 | 2 | 0.880366 | 1 | 0 |
| hf-013 | 2 | 0.755164 | 1 | 0 |
| hf-014 | 2 | 0.826373 | 1 | 0 |
| hf-007 | 2 | 0.267166 | 1 | 0 |
| hf-010 | 2 | 0.955022 | 1 | 0 |
| hf-003 | 2 | 0.24968 | 1 | 0 |
| hf-026 | 2 | 0.6188 | 1 | 0 |
| hf-015 | 2 | 0.197416 | 1 | 0 |
| hf-006 | 2 | 0.561965 | 1 | 0 |
| hf-024 | 2 | 0.382038 | 1 | 0 |
| hf-028 | 2 | 0.506323 | 1 | 0 |
| hf-030 | 2 | 0.378947 | 1 | 0 |
| hf-002 | 2 | 0.322905 | 1 | 0 |
| hf-027 | 2 | 0.15659 | 1 | 0 |
| hf-012 | 2 | 0.116817 | 1 | 0 |
| hf-031 | 3 | 0.952171 | 1 | 0 |
| hf-004 | 3 | 0.374735 | 1 | 0 |
| hf-030 | 3 | 0.378947 | 1 | 0 |
| hf-017 | 3 | 0.880366 | 1 | 0 |
| hf-009 | 3 | 0.197127 | 1 | 0 |
| hf-025 | 3 | 0.139681 | 1 | 0 |
| hf-029 | 3 | 0.91998 | 1 | 0 |
| hf-024 | 3 | 0.382038 | 1 | 0 |
| hf-008 | 3 | 0.276091 | 1 | 0 |
| hf-015 | 3 | 0.197416 | 1 | 0 |
| hf-001 | 3 | 0.987002 | 1 | 0 |
| hf-012 | 3 | 0.116817 | 1 | 0 |
| hf-011 | 3 | 0.239254 | 1 | 0 |
| hf-010 | 3 | 0.955022 | 1 | 0 |
| hf-019 | 3 | 0.26784 | 1 | 0 |
| hf-013 | 3 | 0.755164 | 1 | 0 |
| hf-003 | 3 | 0.24968 | 1 | 0 |
| hf-014 | 3 | 0.826373 | 1 | 0 |
| hf-018 | 3 | 0.925935 | 1 | 0 |
| hf-006 | 3 | 0.561965 | 1 | 0 |
| hf-026 | 3 | 0.6188 | 1 | 0 |
| hf-021 | 3 | 0.303212 | 1 | 0 |
| hf-028 | 3 | 0.506323 | 1 | 0 |
| hf-000 | 3 | 1.005729 | 1 | 0 |
| hf-002 | 3 | 0.322905 | 1 | 0 |
| hf-007 | 3 | 0.267166 | 1 | 0 |
| hf-027 | 3 | 0.15659 | 1 | 0 |
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