heurigym/config.py at main · cornell-zhang/heurigym · GitHub

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
from typing import Dict, TypedDict

class ModelPricing(TypedDict):
    input_price_per_1m: float
    output_price_per_1m: float

# Prices per 1 million tokens
MODEL_PRICING: Dict[str, ModelPricing] = {
    # ─────────────── OpenAI ───────────────
    "gpt-4-turbo-preview": {
        "input_price_per_1m": 10.0,
        "output_price_per_1m": 30.0,
    },
    "gpt-4": {
        "input_price_per_1m": 30.0,
        "output_price_per_1m": 60.0,
    },
    "gpt-3.5-turbo": {
        "input_price_per_1m": 1.0,
        "output_price_per_1m": 2.0,
    },
    "gpt-4o": {
        "input_price_per_1m": 5.0,
        "output_price_per_1m": 15.0,
    },
    "gpt-o3": {
        "input_price_per_1m": 10.0,
        "output_price_per_1m": 40.0,
    },
    "gpt-4.1": {
        "input_price_per_1m": 2.0,
        "output_price_per_1m": 8.0,
    },
    "gpt-4o-mini-high": {
        "input_price_per_1m": 0.15,
        "output_price_per_1m": 0.60,
    },

    # ────────────── Anthropic ─────────────
    "claude-3-opus-20240229": {
        "input_price_per_1m": 15.0,
        "output_price_per_1m": 75.0,
    },
    "claude-3-sonnet-20240229": {
        "input_price_per_1m": 3.0,
        "output_price_per_1m": 15.0,
    },
    "claude-3-haiku-20240307": {
        "input_price_per_1m": 0.25,
        "output_price_per_1m": 1.25,
    },

    # ─────────────── DeepSeek ─────────────
    "deepseek-chat": {
        "input_price_per_1m": 0.27,
        "output_price_per_1m": 1.10,
    },
    "deepseek-reasoner": {
        "input_price_per_1m": 0.55,
        "output_price_per_1m": 2.19,
    },

    # ─────────────── Google ───────────────
    "gemini-1.0-pro": {
        "input_price_per_1m": 0.25,
        "output_price_per_1m": 0.50,
    },
    "gemini-1.5-pro": {
        "input_price_per_1m": 0.50,
        "output_price_per_1m": 1.50,
    },
    "gemini-2.5-flash-preview-04-17": {
        "input_price_per_1m": 0.10,
        "output_price_per_1m": 0.40,
    },
    "gemini-2.5-pro-exp-03-25": {
        "input_price_per_1m": 1.25,
        "output_price_per_1m": 10.0,
    },

    # ──────────────── Meta ────────────────
    "meta-llama/Llama-3.3-70B-Instruct": {
        "input_price_per_1m": 0.80,
        "output_price_per_1m": 0.88,
    },
    "meta-llama/Llama-4-Maverick-17B-128E-Instruct": {
        "input_price_per_1m": 0.27,
        "output_price_per_1m": 0.85,
    },

    # ─────────────── Alibaba ──────────────
    "qwen3-235b-a22b": {
        "input_price_per_1m": 0.20,
        "output_price_per_1m": 0.60,
    },

    # ──────────────── xAI ─────────────────
    "grok-3": {
        "input_price_per_1m": 3.0,
        "output_price_per_1m": 15.0,
    },
}


# Default pricing for unknown models
DEFAULT_PRICING: ModelPricing = {
    "input_price_per_1m": 1.0,
    "output_price_per_1m": 2.0
}

def get_model_pricing(model_name: str) -> ModelPricing:
    """Get pricing information for a specific model."""
    # Handle OpenRouter models by removing the prefix
    if model_name.startswith("openrouter/"):
        model_name = model_name.replace("openrouter/", "", 1)
    return MODEL_PRICING.get(model_name, DEFAULT_PRICING)

def calculate_cost(model_name: str, prompt_tokens: int, completion_tokens: int) -> float:
    """Calculate the cost for a given model and token usage."""
    pricing = get_model_pricing(model_name)
    input_cost = (prompt_tokens * pricing["input_price_per_1m"]) / 1000000
    output_cost = (completion_tokens * pricing["output_price_per_1m"]) / 1000000
    return input_cost + output_cost