atomr-infer-runtime-openai 0.8.0

OpenAI Chat Completions provider for atomr-infer — implements ModelRunner against api.openai.com, with cost table, error classification, and SSE streaming.
Documentation 
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

//! Per-million-token pricing for the OpenAI catalog. Values are public
//! list pricing as of the doc snapshot; operators override via
//! `inference-cli` flags.

use std::collections::HashMap;

use once_cell::sync::Lazy;

#[derive(Debug, Clone, Copy)]
pub struct ModelPrice {
    pub input_per_mtok_usd: f64,
    pub output_per_mtok_usd: f64,
}

pub struct OpenAiPricing {
    table: HashMap<&'static str, ModelPrice>,
}

impl OpenAiPricing {
    pub fn published() -> &'static Self {
        &PRICING
    }

    pub fn get(&self, model: &str) -> Option<ModelPrice> {
        self.table.get(model).copied()
    }
}

static PRICING: Lazy<OpenAiPricing> = Lazy::new(|| {
    let mut t = HashMap::new();
    t.insert(
        "gpt-4o",
        ModelPrice {
            input_per_mtok_usd: 2.50,
            output_per_mtok_usd: 10.00,
        },
    );
    t.insert(
        "gpt-4o-mini",
        ModelPrice {
            input_per_mtok_usd: 0.15,
            output_per_mtok_usd: 0.60,
        },
    );
    t.insert(
        "gpt-4-turbo",
        ModelPrice {
            input_per_mtok_usd: 10.00,
            output_per_mtok_usd: 30.00,
        },
    );
    t.insert(
        "o1-preview",
        ModelPrice {
            input_per_mtok_usd: 15.00,
            output_per_mtok_usd: 60.00,
        },
    );
    t.insert(
        "o1-mini",
        ModelPrice {
            input_per_mtok_usd: 3.00,
            output_per_mtok_usd: 12.00,
        },
    );
    OpenAiPricing { table: t }
});