File size: 6,476 Bytes
18b47fb
 
 
 
7229c6f
18b47fb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7229c6f
18b47fb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7229c6f
18b47fb
 
 
7229c6f
 
 
18b47fb
 
 
 
 
 
7229c6f
18b47fb
7229c6f
18b47fb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7229c6f
 
 
18b47fb
 
 
 
 
 
 
 
 
 
7229c6f
18b47fb
 
 
 
 
 
 
 
 
 
 
 
 
7229c6f
 
 
 
 
18b47fb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7229c6f
 
18b47fb
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
"use client";

import { useState, useMemo } from "react";
import {
  XAxis, YAxis, CartesianGrid,
  Tooltip, ResponsiveContainer, Legend,
  AreaChart, Area,
} from "recharts";

const MODELS = [
  { id: "claude-sonnet", label: "Claude Sonnet 4", inputPer1k: 0.003, outputPer1k: 0.015 },
  { id: "claude-haiku", label: "Claude Haiku 4", inputPer1k: 0.00025, outputPer1k: 0.00125 },
  { id: "gpt-4o-mini", label: "GPT-4o-mini", inputPer1k: 0.00015, outputPer1k: 0.0006 },
  { id: "gpt-4o", label: "GPT-4o", inputPer1k: 0.0025, outputPer1k: 0.01 },
];

export function CostAnalysis() {
  const [numQueries, setNumQueries] = useState(10000);
  const [modelIdx, setModelIdx] = useState(0);
  const model = MODELS[modelIdx];

  const baselineAvgTokens = 950;
  const graphragAvgTokens = 2400;
  const baselineCostPerQ = (800 / 1000) * model.inputPer1k + (150 / 1000) * model.outputPer1k;
  const graphragCostPerQ = (2200 / 1000) * model.inputPer1k + (200 / 1000) * model.outputPer1k;

  const cumulativeData = useMemo(() => {
    const points: { queries: number; Baseline: number; GraphRAG: number }[] = [];
    const step = Math.max(Math.floor(numQueries / 50), 1);
    for (let q = 0; q <= numQueries; q += step) {
      points.push({
        queries: q,
        Baseline: +(baselineCostPerQ * q).toFixed(4),
        GraphRAG: +(graphragCostPerQ * q).toFixed(4),
      });
    }
    return points;
  }, [numQueries, baselineCostPerQ, graphragCostPerQ]);

  return (
    <div>
      {/* Controls */}
      <div className="card mb-6">
        <div className="display-sm mb-4">Cost &amp; Token Projections</div>
        <div className="grid grid-cols-1 md:grid-cols-3 gap-6">
          <div>
            <label className="caption">Number of Queries</label>
            <input type="range" min={100} max={100000} step={100}
              value={numQueries} onChange={(e) => setNumQueries(+e.target.value)}
              className="w-full mt-2 accent-[#FF6B00]" />
            <div className="body-sm font-mono mt-1">{numQueries.toLocaleString()}</div>
          </div>
          <div>
            <label className="caption">LLM Model</label>
            <div className="flex flex-wrap gap-2 mt-2">
              {MODELS.map((m, i) => (
                <button key={m.id}
                  className={i === modelIdx ? "badge-orange" : "badge-outline cursor-pointer"}
                  onClick={() => setModelIdx(i)} style={{ fontSize: "0.75rem" }}>
                  {m.label}
                </button>
              ))}
            </div>
          </div>
          <div>
            <label className="caption">Token Ratio</label>
            <div className="metric-value-sm mt-2" style={{ color: "#cc785c" }}>
              {(graphragAvgTokens / baselineAvgTokens).toFixed(1)}x
            </div>
            <div className="metric-label">GraphRAG / Baseline</div>
          </div>
        </div>
      </div>

      {/* Summary Cards */}
      <div className="grid grid-cols-2 lg:grid-cols-5 gap-4 mb-6">
        {[
          { label: "Cost/Query (Baseline)", value: "$" + baselineCostPerQ.toFixed(6), color: "#0072CE" },
          { label: "Cost/Query (GraphRAG)", value: "$" + graphragCostPerQ.toFixed(6), color: "#FF6B00" },
          { label: `Total (${(numQueries / 1000).toFixed(0)}K)`, value: "$" + (baselineCostPerQ * numQueries).toFixed(2), color: "#0072CE" },
          { label: `Total (${(numQueries / 1000).toFixed(0)}K)`, value: "$" + (graphragCostPerQ * numQueries).toFixed(2), color: "#FF6B00" },
          { label: "Annual (1K qpd)", value: "$" + (graphragCostPerQ * 1000 * 365).toFixed(0), color: "#cc785c" },
        ].map((m, i) => (
          <div key={i} className="card-cream" style={{ padding: "16px", textAlign: "center" }}>
            <div className="metric-value-sm" style={{ color: m.color, fontSize: "1.125rem" }}>{m.value}</div>
            <div className="metric-label">{m.label}</div>
          </div>
        ))}
      </div>

      {/* Cumulative Cost Chart */}
      <div className="card mb-6">
        <div className="title-md mb-4">Cumulative Cost — {model.label}</div>
        <ResponsiveContainer width="100%" height={380}>
          <AreaChart data={cumulativeData} margin={{ top: 10, right: 30, left: 10, bottom: 0 }}>
            <defs>
              <linearGradient id="baselineGrad" x1="0" y1="0" x2="0" y2="1">
                <stop offset="5%" stopColor="#0072CE" stopOpacity={0.2} />
                <stop offset="95%" stopColor="#0072CE" stopOpacity={0} />
              </linearGradient>
              <linearGradient id="graphragGrad" x1="0" y1="0" x2="0" y2="1">
                <stop offset="5%" stopColor="#FF6B00" stopOpacity={0.2} />
                <stop offset="95%" stopColor="#FF6B00" stopOpacity={0} />
              </linearGradient>
            </defs>
            <CartesianGrid strokeDasharray="3 3" stroke="#002B49" strokeOpacity={0.06} />
            <XAxis dataKey="queries" tick={{ fill: "#6c6a64", fontSize: 11 }}
              tickFormatter={(v) => (Number(v) >= 1000 ? `${Number(v) / 1000}K` : String(v))} />
            <YAxis tick={{ fill: "#6c6a64", fontSize: 11 }}
              tickFormatter={(v) => `$${v}`} />
            <Tooltip contentStyle={{ background: "#faf9f5", border: "1px solid #e6dfd8", borderRadius: "8px" }} />
            <Legend />
            <Area type="monotone" dataKey="Baseline" stroke="#0072CE" strokeWidth={2.5} fill="url(#baselineGrad)" />
            <Area type="monotone" dataKey="GraphRAG" stroke="#FF6B00" strokeWidth={2.5} fill="url(#graphragGrad)" />
          </AreaChart>
        </ResponsiveContainer>
      </div>

      {/* Insight Card */}
      <div className="card-coral">
        <div className="display-sm" style={{ color: "white" }}>💡 Key Insight</div>
        <p className="body-md mt-3" style={{ color: "rgba(255,255,255,0.9)", maxWidth: "640px" }}>
          GraphRAG uses <strong>{(graphragAvgTokens / baselineAvgTokens).toFixed(1)}×</strong> more
          tokens per query, but delivers <strong>+13% higher F1</strong> on complex multi-hop questions.
          The Adaptive Router eliminates this overhead for simple queries by routing them to Baseline RAG —
          achieving the best of both worlds.
        </p>
        <button className="btn btn-on-dark mt-4"
          onClick={() => document.getElementById("live")?.scrollIntoView({ behavior: "smooth" })}>
          Try Adaptive Routing →
        </button>
      </div>
    </div>
  );
}