decision-conformance / jev-27b-vl/hf-ba3f0d5/gguf-q3_k_m-iammrduncan/llamacpp-cuda-b11370-jev /

2026-10-09-core-greedy-nothink-adb6a2

Checks: 49/49 passed

Check	Passed	Failures
invariants	33/33	none
batched vs separate	6/6	none
repeat stability	6/6	none
option order	2/2	none
monotonic	2/2	none

Run

Subject: jev-27b-vl/hf-ba3f0d5/gguf-q3_k_m-iammrduncan/llamacpp-cuda-b11370-jev
Suite: decision-conformance · core tier · profile greedy-nothink · mode native
Items: 17 succeeded of 17 attempted (17 planned)
When: 2026-10-09 06:15:24 UTC · 1 min
Hardware: NVIDIA P100 · 16 GB
Harness: bench @ df0f513 · scorer bench/decision-conformance 1
Suite hash: 554e796a858218e48bcb83da003554769782a4a32d3e061da87510d7f9cea490
Recipe: jev-27b-vl/p100/llamacpp · profile gguf-q3_k_m · inference-engines @ 3058df4
Latency: p50 6.51 s · p95 10.60 s · p99 10.60 s (per request)
Usage: 4,776 input tokens · 0 output tokens · $0.00

Reproduce

npm run bench -- run --recipe jev-27b-vl/p100/llamacpp \
    --recipe-profile gguf-q3_k_m \
    --on <host with a p100:p100> \
    --suite decision-conformance --tier core --profile greedy-nothink

Files

exactly as committed under results/