{ "config": { "model": "mistralai/ministral-3-3b-instruct-2512", "host": "172.27.128.1", "port": 1269, "timeoutSec": 60, "maxTokens": 12000, "context": 32768, "outputCap": 12000, "reasoningCounted": true, "quants": "K/V Q4_0" }, "rows": [ { "n": 1, "wallMs": 60011, "totalTokens": 4248, "totalReasoningTokens": 0, "totalAllTokens": 4248, "combinedTokPerSec": 71, "combinedTokPerSecAll": 71, "meanAgentTokPerSec": 174.4, "meanAgentTokPerSecAll": 174.4, "ttftMeanMs": 92, "ttftMaxMs": 92, "ok": 0, "timeout": 1, "errors": 0, "scaleEffPct": 100, "perAgentRelPct": 246 }, { "n": 2, "wallMs": 60033, "totalTokens": 7026, "totalReasoningTokens": 0, "totalAllTokens": 7026, "combinedTokPerSec": 117, "combinedTokPerSecAll": 117, "meanAgentTokPerSec": 116.1, "meanAgentTokPerSecAll": 116.1, "ttftMeanMs": 112, "ttftMaxMs": 125, "ok": 0, "timeout": 2, "errors": 0, "scaleEffPct": 165, "perAgentRelPct": 164 }, { "n": 3, "wallMs": 60040, "totalTokens": 9730, "totalReasoningTokens": 0, "totalAllTokens": 9730, "combinedTokPerSec": 162, "combinedTokPerSecAll": 162, "meanAgentTokPerSec": 77.0, "meanAgentTokPerSecAll": 77.5, "ttftMeanMs": 3526, "ttftMaxMs": 3548, "ok": 0, "timeout": 3, "errors": 0, "scaleEffPct": 228, "perAgentRelPct": 109 }, { "n": 4, "wallMs": 60037, "totalTokens": 12100, "totalReasoningTokens": 0, "totalAllTokens": 12100, "combinedTokPerSec": 202, "combinedTokPerSecAll": 202, "meanAgentTokPerSec": 72.5, "meanAgentTokPerSecAll": 72.6, "ttftMeanMs": 3871, "ttftMaxMs": 3894, "ok": 0, "timeout": 4, "errors": 0, "scaleEffPct": 285, "perAgentRelPct": 102 }, { "n": 5, "wallMs": 60039, "totalTokens": 13500, "totalReasoningTokens": 0, "totalAllTokens": 13500, "combinedTokPerSec": 225, "combinedTokPerSecAll": 225, "meanAgentTokPerSec": 51.1, "meanAgentTokPerSecAll": 51.1, "ttftMeanMs": 6032, "ttftMaxMs": 6059, "ok": 0, "timeout": 5, "errors": 0, "scaleEffPct": 317, "perAgentRelPct": 72 }, { "n": 6, "wallMs": 60039, "totalTokens": 14961, "totalReasoningTokens": 0, "totalAllTokens": 14961, "combinedTokPerSec": 249, "combinedTokPerSecAll": 249, "meanAgentTokPerSec": 48.4, "meanAgentTokPerSecAll": 48.4, "ttftMeanMs": 6961, "ttftMaxMs": 6988, "ok": 0, "timeout": 6, "errors": 0, "scaleEffPct": 351, "perAgentRelPct": 68 }, { "n": 7, "wallMs": 60039, "totalTokens": 16443, "totalReasoningTokens": 0, "totalAllTokens": 16443, "combinedTokPerSec": 274, "combinedTokPerSecAll": 274, "meanAgentTokPerSec": 46.7, "meanAgentTokPerSecAll": 46.7, "ttftMeanMs": 7570, "ttftMaxMs": 7594, "ok": 0, "timeout": 7, "errors": 0, "scaleEffPct": 386, "perAgentRelPct": 66 }, { "n": 8, "wallMs": 60043, "totalTokens": 17130, "totalReasoningTokens": 0, "totalAllTokens": 17130, "combinedTokPerSec": 285, "combinedTokPerSecAll": 285, "meanAgentTokPerSec": 41.8, "meanAgentTokPerSecAll": 41.8, "ttftMeanMs": 8694, "ttftMaxMs": 8718, "ok": 0, "timeout": 8, "errors": 0, "scaleEffPct": 401, "perAgentRelPct": 59 }, { "n": 9, "wallMs": 60039, "totalTokens": 18175, "totalReasoningTokens": 0, "totalAllTokens": 18175, "combinedTokPerSec": 303, "combinedTokPerSecAll": 303, "meanAgentTokPerSec": 40.6, "meanAgentTokPerSecAll": 40.6, "ttftMeanMs": 9202, "ttftMaxMs": 9230, "ok": 0, "timeout": 9, "errors": 0, "scaleEffPct": 427, "perAgentRelPct": 57 }, { "n": 10, "wallMs": 60039, "totalTokens": 19413, "totalReasoningTokens": 0, "totalAllTokens": 19413, "combinedTokPerSec": 323, "combinedTokPerSecAll": 323, "meanAgentTokPerSec": 38.6, "meanAgentTokPerSecAll": 38.6, "ttftMeanMs": 9667, "ttftMaxMs": 9690, "ok": 0, "timeout": 10, "errors": 0, "scaleEffPct": 455, "perAgentRelPct": 54 }, { "n": 11, "wallMs": 60043, "totalTokens": 19277, "totalReasoningTokens": 0, "totalAllTokens": 19277, "combinedTokPerSec": 321, "combinedTokPerSecAll": 321, "meanAgentTokPerSec": 38.4, "meanAgentTokPerSecAll": 38.4, "ttftMeanMs": 11420, "ttftMaxMs": 11450, "ok": 0, "timeout": 11, "errors": 0, "scaleEffPct": 452, "perAgentRelPct": 54 }, { "n": 12, "wallMs": 60042, "totalTokens": 21114, "totalReasoningTokens": 0, "totalAllTokens": 21114, "combinedTokPerSec": 352, "combinedTokPerSecAll": 352, "meanAgentTokPerSec": 37.3, "meanAgentTokPerSecAll": 37.3, "ttftMeanMs": 11637, "ttftMaxMs": 11668, "ok": 0, "timeout": 12, "errors": 0, "scaleEffPct": 496, "perAgentRelPct": 53 }, { "n": 13, "wallMs": 60047, "totalTokens": 21976, "totalReasoningTokens": 0, "totalAllTokens": 21976, "combinedTokPerSec": 366, "combinedTokPerSecAll": 366, "meanAgentTokPerSec": 36.1, "meanAgentTokPerSecAll": 36.1, "ttftMeanMs": 12076, "ttftMaxMs": 12109, "ok": 0, "timeout": 13, "errors": 0, "scaleEffPct": 515, "perAgentRelPct": 51 }, { "n": 14, "wallMs": 60046, "totalTokens": 22169, "totalReasoningTokens": 0, "totalAllTokens": 22169, "combinedTokPerSec": 369, "combinedTokPerSecAll": 369, "meanAgentTokPerSec": 33.6, "meanAgentTokPerSecAll": 33.6, "ttftMeanMs": 12561, "ttftMaxMs": 12593, "ok": 0, "timeout": 14, "errors": 0, "scaleEffPct": 520, "perAgentRelPct": 47 }, { "n": 15, "wallMs": 60044, "totalTokens": 23092, "totalReasoningTokens": 0, "totalAllTokens": 23092, "combinedTokPerSec": 385, "combinedTokPerSecAll": 385, "meanAgentTokPerSec": 32.6, "meanAgentTokPerSecAll": 32.6, "ttftMeanMs": 12833, "ttftMaxMs": 12871, "ok": 0, "timeout": 15, "errors": 0, "scaleEffPct": 542, "perAgentRelPct": 46 }, { "n": 16, "wallMs": 60044, "totalTokens": 23822, "totalReasoningTokens": 0, "totalAllTokens": 23822, "combinedTokPerSec": 397, "combinedTokPerSecAll": 397, "meanAgentTokPerSec": 32.0, "meanAgentTokPerSecAll": 32.0, "ttftMeanMs": 13437, "ttftMaxMs": 13505, "ok": 0, "timeout": 16, "errors": 0, "scaleEffPct": 559, "perAgentRelPct": 45 }, { "n": 17, "wallMs": 60052, "totalTokens": 22402, "totalReasoningTokens": 0, "totalAllTokens": 22402, "combinedTokPerSec": 373, "combinedTokPerSecAll": 373, "meanAgentTokPerSec": 28.7, "meanAgentTokPerSecAll": 28.7, "ttftMeanMs": 13973, "ttftMaxMs": 14061, "ok": 0, "timeout": 17, "errors": 0, "scaleEffPct": 525, "perAgentRelPct": 40 }, { "n": 18, "wallMs": 60044, "totalTokens": 23572, "totalReasoningTokens": 0, "totalAllTokens": 23572, "combinedTokPerSec": 393, "combinedTokPerSecAll": 393, "meanAgentTokPerSec": 27.5, "meanAgentTokPerSecAll": 27.5, "ttftMeanMs": 12412, "ttftMaxMs": 12524, "ok": 0, "timeout": 18, "errors": 0, "scaleEffPct": 554, "perAgentRelPct": 39 }, { "n": 19, "wallMs": 60044, "totalTokens": 23481, "totalReasoningTokens": 0, "totalAllTokens": 23481, "combinedTokPerSec": 391, "combinedTokPerSecAll": 391, "meanAgentTokPerSec": 26.5, "meanAgentTokPerSecAll": 26.5, "ttftMeanMs": 13272, "ttftMaxMs": 13427, "ok": 0, "timeout": 19, "errors": 0, "scaleEffPct": 551, "perAgentRelPct": 37 }, { "n": 20, "wallMs": 60046, "totalTokens": 23724, "totalReasoningTokens": 0, "totalAllTokens": 23724, "combinedTokPerSec": 395, "combinedTokPerSecAll": 395, "meanAgentTokPerSec": 26.2, "meanAgentTokPerSecAll": 26.2, "ttftMeanMs": 14655, "ttftMaxMs": 14791, "ok": 0, "timeout": 20, "errors": 0, "scaleEffPct": 556, "perAgentRelPct": 37 }, { "n": 21, "wallMs": 60047, "totalTokens": 23393, "totalReasoningTokens": 0, "totalAllTokens": 23393, "combinedTokPerSec": 390, "combinedTokPerSecAll": 390, "meanAgentTokPerSec": 25.7, "meanAgentTokPerSecAll": 25.7, "ttftMeanMs": 16613, "ttftMaxMs": 16764, "ok": 0, "timeout": 21, "errors": 0, "scaleEffPct": 549, "perAgentRelPct": 36 }, { "n": 22, "wallMs": 60048, "totalTokens": 24238, "totalReasoningTokens": 0, "totalAllTokens": 24238, "combinedTokPerSec": 404, "combinedTokPerSecAll": 404, "meanAgentTokPerSec": 24.8, "meanAgentTokPerSecAll": 24.8, "ttftMeanMs": 15568, "ttftMaxMs": 15737, "ok": 0, "timeout": 22, "errors": 0, "scaleEffPct": 569, "perAgentRelPct": 35 }, { "n": 23, "wallMs": 60050, "totalTokens": 23529, "totalReasoningTokens": 0, "totalAllTokens": 23529, "combinedTokPerSec": 392, "combinedTokPerSecAll": 392, "meanAgentTokPerSec": 23.4, "meanAgentTokPerSecAll": 23.4, "ttftMeanMs": 16333, "ttftMaxMs": 16542, "ok": 0, "timeout": 23, "errors": 0, "scaleEffPct": 552, "perAgentRelPct": 33 }, { "n": 24, "wallMs": 60050, "totalTokens": 24026, "totalReasoningTokens": 0, "totalAllTokens": 24026, "combinedTokPerSec": 400, "combinedTokPerSecAll": 400, "meanAgentTokPerSec": 23.1, "meanAgentTokPerSecAll": 23.1, "ttftMeanMs": 16549, "ttftMaxMs": 16771, "ok": 0, "timeout": 24, "errors": 0, "scaleEffPct": 563, "perAgentRelPct": 33 } ] }