{ "ts": 1785306125.3880131, "host": "dgx-spark", "engine": "poolsideai/llama.cpp 04b2b72", "model": "laguna-s-2.1-Q4_K_M.gguf", "sha256": "a8b55c75714ea73fd90ec85de5defdc0b8d88ca0ad2108343cdd8fc22f7583e4", "serve": { "host": "0.0.0.0", "port": 8000, "ctx": 8192, "ngl": -1, "parallel": 1, "alias": "local-laguna", "jinja": true, "fa": "on" }, "throughput": [ { "prompt_tokens": 46, "completion_tokens": 4, "latency_s": 0.377, "tok_s": 10.612, "label": "gen8_short" }, { "prompt_tokens": 67, "completion_tokens": 128, "latency_s": 5.962, "tok_s": 21.47, "label": "gen128" }, { "prompt_tokens": 59, "completion_tokens": 247, "latency_s": 11.288, "tok_s": 21.881, "label": "gen256" } ], "quote_gen_tok_s": 21.47, "gen128_pre_smoke": { "label": "gen128", "prompt_tokens": 67, "completion_tokens": 128, "latency_s": 5.69, "tok_s": 22.497, "temp": 0.0, "model": "local-laguna", "ts": 1785305130.4795225 }, "agent_smoke": { "pass": "40/40", "elapsed_s": 84.86, "temperature": 0.0, "model": "local-laguna", "base_url": "http://127.0.0.1:8000/v1", "host": "dgx-spark", "out": "results/agent_smoke.json", "runner_sha256": "3bb81080879ddf78d3d6295e333c78b1ecc1d84ba5a9e9dc19a0e677f54da48f", "run_manifest_schema": "laguna.run_manifest/v1", "lock_note": "accel multi-bench 2026-07-29T13:22:05.385271+07:00" }, "hermes_agent_smoke": { "pass": "27/27", "elapsed_s": 100.1, "temperature": 0.0, "model": "local-laguna", "base_url": "http://127.0.0.1:8000/v1", "host": "dgx-spark", "out": "results/hermes_agent_smoke.json", "runner_sha256": "20c1e52a8a22306b81309c08984d6875cfff46e7382a46b674c1a6345030293b", "run_manifest_schema": "laguna.run_manifest/v1", "protocol": "one-response protocol; tools validated not executed", "lock_note": "accel multi-bench 2026-07-29T13:22:05.385271+07:00" }, "server_bench_live": { "base_url": "http://127.0.0.1:8000/v1", "model": "local-laguna", "rows": [ { "mark": "2k", "latency_s": 1.597, "prompt_tokens": 836, "completion_tokens": 3, "completion_tok_s": 1.878, "content": "OK" }, { "mark": "8k", "latency_s": 4.78, "prompt_tokens": 3236, "completion_tokens": 3, "completion_tok_s": 0.628, "content": "OK" } ], "ts": 1785305345.2721028 }, "git_head": "bf82eab5fd6c1fb04e863f0c4b05b5658dec4aee", "git_status_line": "## main...origin/main", "schema": "laguna.measured/v1", "updated_at_wib": "2026-07-29T13:22:05.385271+07:00", "headline_quant": "Q4_K_M", "headline": "40/40 agent_smoke \u00b7 official Q4_K_M \u00b7 hermes 27/27 \u00b7 gen128 ~21.47 t/s", "accel_run_id": "20260729T132205+0700" }