Cost-Performance — Animate | AI Wisdom

← Back to topic
🎞 ANIMATE · Mode 3 of 5 · +100 XP— LLM Cost vs Performance
📋Pricing
⚙️Calculate
🔄Compare
✅Optimize
🎯 Next: Test →
Step 1 / 4
Speed
Pricing Models
All providers charge per token — input and output priced separately💡 Analogy
Like a printing service: cheaper to give them paper (input) than to buy printed pages (output).
LLM pricing: $/1M input tokens and $/1M output tokens. Input is always cheaper than output. Gemini Flash is the cheapest; GPT-4o and Claude Sonnet are premium.
Console Output
> Cheapest: gemini-1.5-flash ($0.075 input)
> Most expensive: gpt-4o ($5.00 input)
> Ratio: 67× price difference
> Output always > input in price
●cost_performance.py
Pricing ModelsL1–8
1# Model cost comparison (May 2025)
2MODELS = {
3    "gpt-4o-mini":          {"input": 0.15,  "output": 0.60},
4    "claude-3-5-haiku":     {"input": 0.25,  "output": 1.25},
5    "gpt-4o":               {"input": 5.00,  "output": 15.00},
6    "claude-3-5-sonnet":    {"input": 3.00,  "output": 15.00},
7    "gemini-1.5-flash":     {"input": 0.075, "output": 0.30},
8    "gemini-1.5-pro":       {"input": 3.50,  "output": 10.50},
9}
10 
11def monthly_cost(model: str, daily_requests: int,
12                 avg_input: int, avg_output: int) -> float:
13    pricing = MODELS[model]
14    per_request = (avg_input * pricing["input"] +
15                   avg_output * pricing["output"]) / 1_000_000
16    return per_request * daily_requests * 30
17 
18# Compare: 100k requests/day, 500 in / 200 out tokens avg
19print("Monthly costs for 100k req/day, 500 in / 200 out tokens:")
20for model, _ in MODELS.items():
21    cost = monthly_cost(model, 100_000, 500, 200)
22    print(f"  {model:<25} ${cost:>8,.2f}/month")
23 
24# Output:
25# gpt-4o-mini               $   645.00/month
26# claude-3-5-haiku          $ 1,575.00/month
27# gpt-4o                    $31,500.00/month
28# claude-3-5-sonnet         $24,000.00/month
29# gemini-1.5-flash          $   337.50/month
30# gemini-1.5-pro            $26,250.00/month
30 lines
Timeline — 4 Steps
Pricing Models
All providers charge per token — input and output priced separately💡 Analogy
Like a printing service: cheaper to give them paper (input) than to buy printed pages (output).
LLM pricing: $/1M input tokens and $/1M output tokens. Input is always cheaper than output. Gemini Flash is the cheapest; GPT-4o and Claude Sonnet are premium.
Console Output
> Cheapest: gemini-1.5-flash ($0.075 input)
> Most expensive: gpt-4o ($5.00 input)
> Ratio: 67× price difference
> Output always > input in price

Was this page helpful?

Discussion

All providers charge per token — input and output priced separately

💡 Analogy

Like a printing service: cheaper to give them paper (input) than to buy printed pages (output).

LLM pricing: $/1M input tokens and $/1M output tokens. Input is always cheaper than output. Gemini Flash is the cheapest; GPT-4o and Claude Sonnet are premium.

Console Output

> Cheapest: gemini-1.5-flash ($0.075 input)
> Most expensive: gpt-4o ($5.00 input)
> Ratio: 67× price difference
> Output always > input in price

All providers charge per token — input and output priced separately

💡 Analogy

Like a printing service: cheaper to give them paper (input) than to buy printed pages (output).

LLM pricing: $/1M input tokens and $/1M output tokens. Input is always cheaper than output. Gemini Flash is the cheapest; GPT-4o and Claude Sonnet are premium.

Console Output

> Cheapest: gemini-1.5-flash ($0.075 input)
> Most expensive: gpt-4o ($5.00 input)
> Ratio: 67× price difference
> Output always > input in price