rakka-inference-runtime-gemini 0.2.6

Google Gemini provider for rakka-inference — implements ModelRunner against Vertex AI and AI Studio endpoints, 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

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 GeminiPricing {
    table: HashMap<&'static str, ModelPrice>,
}

impl GeminiPricing {
    pub fn published() -> &'static Self {
        &PRICING
    }
    pub fn get(&self, model: &str) -> Option<ModelPrice> {
        self.table.get(model).copied()
    }
}

static PRICING: Lazy<GeminiPricing> = Lazy::new(|| {
    let mut t = HashMap::new();
    t.insert(
        "gemini-2.0-pro",
        ModelPrice {
            input_per_mtok_usd: 1.25,
            output_per_mtok_usd: 5.0,
        },
    );
    t.insert(
        "gemini-1.5-pro",
        ModelPrice {
            input_per_mtok_usd: 1.25,
            output_per_mtok_usd: 5.0,
        },
    );
    t.insert(
        "gemini-1.5-flash",
        ModelPrice {
            input_per_mtok_usd: 0.075,
            output_per_mtok_usd: 0.30,
        },
    );
    t.insert(
        "gemini-2.0-flash",
        ModelPrice {
            input_per_mtok_usd: 0.075,
            output_per_mtok_usd: 0.30,
        },
    );
    GeminiPricing { table: t }
});