NVIDIA NIM

Inference81 modelcontext maks 1M

Catatan dari sumber

Free with NVIDIA Developer Program membership. 100+ models. Rate-limited per model.

Setup dalam 5 menit

Model

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://integrate.api.nvidia.com/v1",
  apiKey: process.env.NVIDIA_NIM_API_KEY,
});

const response = await client.chat.completions.create({
  model: "deepseek-ai-deepseek-v4-flash",
  messages: [{ role: "user", content: "Halo!" }],
});

console.log(response.choices[0].message.content);

.env

NVIDIA_NIM_API_KEY=your-api-key-here

Belum punya API key? Ambil di halaman resmi NVIDIA NIM ↗.

Cara claim API key gratis

Langkah umum — detail pastinya ikutin halaman resmi NVIDIA NIM.

1.Buka halaman API key NVIDIA NIM ↗
2.Daftar akun baru, atau login kalau udah punya.
3.Generate API key di dashboard / settings.
4.Pakai API key + Base URL https://integrate.api.nvidia.com/v1 di SDK atau HTTP client.

Model tersedia (81+)

deepseek-ai/deepseek-v4-flash

deepseek-ai-deepseek-v4-flash

Modality: text + reasoning
Context: 1.0M
Rate limit: Up to 40 RPM

deepseek-ai/deepseek-v4-pro

deepseek-ai-deepseek-v4-pro

Modality: text + reasoning
Context: 1.0M
Rate limit: Up to 40 RPM

Kimi K3

kimi-k3

Modality: vision + reasoning
Context: 1.0M

meta/llama-guard-4-12b

meta-llama-guard-4-12b

Modality: text + image
Context: 1.0M
Output: 16K
Rate limit: Up to 40 RPM

minimaxai/minimax-m3

Modality: Text
Context: 1M
Output: ~64K
Rate limit: 40 RPM, 10,000 RPD

nvidia/nemotron-3-super-120b-a12b

Modality: Text
Context: 1M
Output: 262K
Rate limit: 40 RPM, 10,000 RPD

nvidia/nemotron-3-ultra-550b-a55b

Modality: Text
Context: 1M
Output: 262K
Rate limit: 40 RPM, 10,000 RPD

google/gemma-4-31b-it

Modality: Text
Context: 262K
Output: 8K
Rate limit: 40 RPM, 10,000 RPD

Model	Modality	Context	Output	Rate limit
deepseek-ai/deepseek-v4-flash deepseek-ai-deepseek-v4-flash	text + reasoning	1.0M		Up to 40 RPM
deepseek-ai/deepseek-v4-pro deepseek-ai-deepseek-v4-pro	text + reasoning	1.0M		Up to 40 RPM
Kimi K3 kimi-k3	vision + reasoning	1.0M
meta/llama-guard-4-12b meta-llama-guard-4-12b	text + image	1.0M	16K	Up to 40 RPM
minimaxai/minimax-m3 minimaxai/minimax-m3	Text	1M	~64K	40 RPM, 10,000 RPD
nvidia/nemotron-3-super-120b-a12b nvidia/nemotron-3-super-120b-a12b	Text	1M	262K	40 RPM, 10,000 RPD
nvidia/nemotron-3-ultra-550b-a55b nvidia/nemotron-3-ultra-550b-a55b	Text	1M	262K	40 RPM, 10,000 RPD
google/gemma-4-31b-it google/gemma-4-31b-it	Text	262K	8K	40 RPM, 10,000 RPD

+ 92 more models model lainnya — lihat daftar lengkap di sumber.

Model ini di provider lain (17)

Nama model yang sama juga muncul di provider lain menurut sumber masing-masing — versi/kuantisasi bisa beda.

FAQ

Apakah API NVIDIA NIM gratis?

Ya — menurut sumber: "Free with NVIDIA Developer Program membership. 100+ models. Rate-limited per model." Data dari mnfst/awesome-free-llm-apis, freellm.net, models.dev, terakhir di-sync 2 Sep 2026 — cek sumber untuk kondisi terbaru.

Ada berapa model gratis di NVIDIA NIM?

81 model tercantum per 2 Sep 2026, termasuk deepseek-ai/deepseek-v4-flash, deepseek-ai/deepseek-v4-pro, Kimi K3.

Berapa rate limit NVIDIA NIM?

Beda-beda per model, contoh: deepseek-ai/deepseek-v4-flash: Up to 40 RPM; deepseek-ai/deepseek-v4-pro: Up to 40 RPM; meta/llama-guard-4-12b: Up to 40 RPM. Limit lengkap per model, lihat tabel di atas.

Berapa context window maksimal di NVIDIA NIM?

1M (minimaxai/minimax-m3).