NVIDIA: Nemotron 3.5 Lightning

Характеристики

Контекстное окно: 262K токенов
Провайдер: 262144
Архитектура: text->text

Модальности

Входящие

text

Исходящие

text

Цены

Оплата по факту за использованные токены и операции. Цены в рублях (за 1 млн токенов или за единицу — как указано).

Тип	Стоимость	Описание
Входящие токены	10.30 ₽ / 1M	Токены запроса и контекста, обработанные моделью
Исходящие токены	25.76 ₽ / 1M	Токены ответа, сгенерированные моделью
Чтение кэша	10.00 ₽ / 1M	Повторное использование кэшированного контекста

Цены могут меняться в зависимости от курса валют. Данные по тарифам поступают от поставщика услуг и обновляются каждые сутки.

Начать использовать

Описание

Here's a thinking process:

Analyze User Request:

Task: Translate a model description from English to Russian
Source Text: "NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total. It is suited for high-throughput agentic workloads and specialized tasks that..."
Constraints:
- Keep model names and technical terms in the "familiar for Russian-speaking developers" style
- Output only the text translation, no introduction, no quotes around the whole text
Input Format: The user provided the model name and the original English text, with a prompt to translate only the description part, keeping names/terms as Russians developers would know them.

Identify Key Elements to Translate/Preserve:

"NVIDIA Nemotron 3.5 Lightning" - This is a model name. In Russian developer context, it's typically kept as-is or slightly adapted. Usually, model names like Nemotron are transliterated or kept in English. The prompt says "Keep names and technical terms in the familiar for Russian-speaking developers view." So I should keep "NVIDIA Nemotron 3.5 Lightning" as is, or maybe "NVIDIA Nemotron 3.5 Лайтнинг" but typically English names are kept. I'll keep the name in the standard Russian dev community format, which usually means keeping English proper nouns as-is, maybe with transliteration if common. "Nemotron" is often kept as "Немоторон" in some contexts, but in dev/tech, English is standard. I'll keep it as "NVIDIA Nemotron 3.5 Lightning" but if translation is needed, I'd consider "NVIDIA Nemotron 3.5 Лайтнинг". Let's see the original: "NVIDIA: Nemotron 3.5 Lightning (nvidia/nemotron-3.5-lightning)". The prompt says "Save model names and technical terms in the familiar for Russian-speaking developers view." I think keeping the English name is fine, but I can add a Russian transliteration in parentheses if needed. However, the instruction says "Keep names... in familiar video", which usually means don't over-translate, keep English terms as they are used in Ru dev community. I'll keep "NVIDIA Nemotron 3.5 Lightning" as is.
"open mixture-of-experts model" -> "открытая модель смешанных экспертов" or "открытая модель с архитектурой Mixture-of-Experts". Technical term "Mixture-of-Experts" is often kept as "Mixture-of-Experts" or translated as "смесь экспертов". In Russian dev community, "Mixture-of-Experts" is often transliterated as "Миксс-оф-Экспертов" or just kept in English. I'll use "смесь экспертов" or keep "Mixture-of-Experts". The prompt says "technical terms in familiar for Russian-speaking developers view". I think "Mixture-of-Experts" is widely recognized, but I can translate it as "архитектура Mixture-of-Experts" or "смесь экспертов". I'll go with a natural Russian technical phrasing: "открытая модель архитектуры Mixture-of-Experts" or just "открытая модель смеси экспертов". Let's see what's standard. Often, it's "модель Mixture-of-Experts (MoE)". I'll write "открытая модель Mixture-of-Experts".
"3B active parameters out of 30B total" -> "3 активных параметра из 30 всего" or "3 миллиарда активных параметров из 30 миллиардов всего". "Parameters" is "параметры". "3B" = "3 млрд" or "3B". I'll write "3 млрд активных параметров из 30 млрд всего".
"high-throughput agentic workloads" -> "высокопроизводительные агентские нагрузки" or "нагрузки высокой пропускной способности для агентов". "Agentic" relates to "агенты" or "агентская". I'll use "агентские нагрузки" or "нагрузки для автономных агентов". Common in Ru dev: "агентские задачи" or "нагрузки для агентов". I'll phrase it as "высокопроизводительные агентские нагрузки".
"specialized tasks that..." -> "специализированные задачи, для которых..." The text ends with "that...", so I need to complete it or just translate what's given. The user gave the full prompt text ending with "that...", but the actual original text in the prompt is: "NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total. It is suited for high-throughput agentic workloads and specialized tasks that..." It seems cut off. I should translate the given part, and if it ends with "that...", I might need to handle the ellipsis. The user says: "Переведи на русский язык описание модели ИИ из каталога." and provides the text. The text ends with "that...". I'll translate up to the end of the provided sentence, or assume it continues. I'll translate the full given sentence, and if it's incomplete, I'll keep the ellipsis or finish naturally. Looking at the prompt: "NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total. It is suited for high-throughput agentic workloads and specialized tasks that..." I'll translate the two sentences, and for the trailing "that...", I'll either keep "that..." or translate it as "которые..." but since it's c…

API и примеры кода

Единый API совместим с OpenAI SDK. Используйте ключ из личного кабинета.

main.py

from openai import OpenAI

client = OpenAI(
    base_url="https://routerapi.ru/api/v1",
    api_key="YOUR_API_KEY",
)

response = client.chat.completions.create(
    model="nvidia/nemotron-3.5-lightning",
    messages=[{"role": "user", "content": "Привет!"}],
)

print(response.choices[0].message.content)

terminal

curl -X POST "https://routerapi.ru/api/v1/chat/completions" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "nvidia/nemotron-3.5-lightning",
    "messages": [
      {"role": "user", "content": "Привет!"}
    ]
  }'

request.php

<?php
declare(strict_types=1);

$response = (new \GuzzleHttp\Client())->post(
    'https://routerapi.ru/api/v1/chat/completions',
    [
        'headers' => [
            'Authorization' => 'Bearer YOUR_API_KEY',
            'Content-Type'  => 'application/json',
        ],
        'json' => [
            'model'    => 'nvidia/nemotron-3.5-lightning',
            'messages' => [
                ['role' => 'user', 'content' => 'Привет!'],
            ],
        ],
    ]
);

$data = json_decode((string) $response->getBody(), true);
echo $data['choices'][0]['message']['content'];

main.go

package main

import (
    "bytes"
    "encoding/json"
    "fmt"
    "io"
    "net/http"
)

func main() {
    body, _ := json.Marshal(map[string]any{
        "model": "nvidia/nemotron-3.5-lightning",
        "messages": []map[string]string{
            {"role": "user", "content": "Привет!"},
        },
    })

    req, _ := http.NewRequest("POST",
        "https://routerapi.ru/api/v1/chat/completions",
        bytes.NewReader(body))
    req.Header.Set("Authorization", "Bearer YOUR_API_KEY")
    req.Header.Set("Content-Type", "application/json")

    resp, _ := http.DefaultClient.Do(req)
    defer resp.Body.Close()

    var result map[string]any
    raw, _ := io.ReadAll(resp.Body)
    json.Unmarshal(raw, &result)

    choices := result["choices"].([]any)
    msg := choices[0].(map[string]any)["message"].(map[string]any)
    fmt.Println(msg["content"])
}

Совместимо с OpenAI Python SDK, Node.js SDK, curl и любым HTTP-клиентом. Документация API · Все модели