Round 6 breakdown

disqualifiedleaderSubmission

Round score

—

degenerate output (baseline-relative repetition failures exceed allowance of 4 prompts (hf-029: distinct 16-gram ratio 0.537 is more than 0.10 below baseline 0.652 over 20210 chars))

Prompts scored

0 / 0

median_e2e_speedup

Below tolerance

answered too little to score

Zeroed total

scored 0.0 without measuring

Workload

Temperature range

0.1 to 1.01 (per prompt)

Request interval

2 ms

Thinking

Disabled

Context limit

262,144 tokens

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 (baseline-relative repetition failures exceed allowance of 4 prompts (hf-029: distinct 16-gram ratio 0.537 is more than 0.10 below baseline 0.652 over 20210 chars))

Request trace

32 requests

Request	TTFT	Total	Tokens	Slowest gap	Stream
hf-000	250.43ms	31.292s	4694	38.25ms	—
hf-001	267.11ms	31.304s	4617	25.18ms	—
hf-002	264.01ms	34.948s

Per-prompt repetition checks

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	0.8903	0.8862	-0.0041	0.1000	Passed
hf-001	0.9233	0.9211	-0.0022	0.1000	Passed
hf-002	0.9477	0.8113	-0.1363	0.1000	Passed
hf-003	0.9768

Prompt	Repetition	Temperature	Top p
hf-028	1	0.966458	1
hf-025	1	0.304724	1
hf-004	1	0.239233	1
hf-018	1	0.635485	1
hf-006	1	0.527097	1
hf-008	1	0.383346	1
hf-029	1	0.240187	1
hf-010	1	0.703177	1
hf-031	1	0.60416	1
hf-000	1	0.261035	1
hf-001	1	0.213557	1
hf-014	1	0.829549	1
hf-017	1	0.169615	1
hf-022	1	0.235223	1
hf-015	1	0.412126	1
hf-020	1	0.377109	1
hf-013	1	0.74317	1
hf-007	1	0.125605	1
hf-027	1	0.461129	1
hf-012	1	0.794148	1
hf-030	1	0.246463	1
hf-024	1	0.101954	1
hf-009	1	0.296916	1
hf-023	1	0.790024	1
hf-016	1	0.116507	1
hf-002	1	0.107098	1
hf-019	1	0.223606	1
hf-026	1	0.910684	1
hf-011	1	0.928575	1
hf-003	1	0.687747	1
hf-005	1	0.786804	1
hf-021	1	0.883976	1
hf-028	2	0.966458	1
hf-025	2	0.304724	1
hf-004	2	0.239233	1
hf-018	2	0.635485	1
hf-006	2	0.527097	1
hf-008	2	0.383346	1
hf-029	2	0.240187	1
hf-010	2	0.703177	1
hf-031	2	0.60416	1
hf-000	2	0.261035	1
hf-001	2	0.213557	1
hf-017	2	0.169615	1
hf-014	2	0.829549	1
hf-022	2	0.235223	1
hf-015	2	0.412126	1
hf-020	2	0.377109	1
hf-013	2	0.74317	1
hf-007	2	0.125605	1
hf-027	2	0.461129	1
hf-012	2	0.794148	1
hf-030	2	0.246463	1
hf-024	2	0.101954	1
hf-009	2	0.296916	1
hf-023	2	0.790024	1
hf-016	2	0.116507	1
hf-002	2	0.107098	1
hf-019	2	0.223606	1
hf-026	2	0.910684	1
hf-011	2	0.928575	1
hf-003	2	0.687747	1
hf-005	2	0.786804	1
hf-021	2	0.883976	1
hf-028	3	0.966458	1
hf-025	3	0.304724	1
hf-004	3	0.239233	1
hf-018	3	0.635485	1
hf-006	3	0.527097	1
hf-008	3	0.383346	1
hf-029	3	0.240187	1
hf-010	3	0.703177	1
hf-031	3	0.60416	1
hf-000	3	0.261035	1
hf-001	3	0.213557	1
hf-017	3	0.169615	1
hf-014	3	0.829549	1
hf-022	3	0.235223	1
hf-015	3	0.412126	1
hf-020	3	0.377109	1
hf-013	3	0.74317	1
hf-007	3	0.125605	1
hf-027	3	0.461129	1
hf-012	3	0.794148	1
hf-030	3	0.246463	1
hf-024	3	0.101954	1
hf-009	3	0.296916	1
hf-023	3	0.790024	1
hf-016	3	0.116507	1
hf-002	3	0.107098	1
hf-019	3	0.223606	1
hf-026	3	0.910684	1
hf-011	3	0.928575	1
hf-003	3	0.687747	1
hf-005	3	0.786804	1
hf-021	3	0.883976	1

Round 6 breakdown

Workload

Request trace

Per-prompt repetition checks

Round 6 breakdown

Round 6 breakdown

Workload

Request trace

Per-prompt repetition checks

Replay sampling

Correctness

SLA replay

Engine