Core Services

Note: This section covers the Inference API (OpenAI compatible), which provides serverless endpoints for running generative AI models without managing infrastructure. Find guides, supported models, and usage examples below.

Inference API (Inference Models)

Welcome to MegaNova's Inference API, the easiest way to integrate powerful generative AI models into your applications. The Inference API is OpenAI compatible and available at https://api.meganova.ai/v1.

With MegaNova, endpoints are pre-configured for you—simply sign up, log in, and start using them right away. You can explore the capabilities of our models live on our platform UI or integrate them into your projects using provided code samples in curl, Python, or JavaScript.

Key Features

OpenAI Compatible: Use OpenAI SDKs and tools with MegaNova's Inference API.
Inference Models: No setup required—get started instantly after signing in.
Live Use: Test models interactively on our platform's UI.
Code Samples: Quick-start guides for curl, Python, and JavaScript.
Broad Model Coverage: Claude, GPT-5 family, Gemini, Grok, Qwen, MiniMax, MiMo, GLM, Kimi, DeepSeek, Llama, Mistral, and more.
Scalable Performance: Built on MegaNova's GPU-accelerated infrastructure.

Prerequisites

MegaNova Account: Ensure you have an account on our platform.
Credit Balance: For paid models, ensure you have credit balance in your account (or enable paid fallback after free quota).
Free Models: No credit balance required for free models (see Free Models section below).

Quickstart with the OpenAI SDK

Because the Inference API is OpenAI-compatible, you can use the official OpenAI SDKs by just overriding base_url and api_key.

Python

from openai import OpenAI

client = OpenAI(
    base_url="https://api.meganova.ai/v1",
    api_key="sk-...",  # your MegaNova API key
)

resp = client.chat.completions.create(
    model="deepseek-ai/DeepSeek-V3.2",
    messages=[{"role": "user", "content": "Say hello in one sentence."}],
)
print(resp.choices[0].message.content)

JavaScript / TypeScript

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.meganova.ai/v1",
  apiKey: process.env.MEGANOVA_API_KEY,
});

const resp = await client.chat.completions.create({
  model: "deepseek-ai/DeepSeek-V3.2",
  messages: [{ role: "user", content: "Say hello in one sentence." }],
});
console.log(resp.choices[0].message.content);

Streaming, function/tool calls, images (client.images.generate), and embeddings (client.embeddings.create) all work through the same client. See Text Generation for more.

Tier-Based Model Access

For the complete, up-to-date catalog with prices and model IDs, see the Model List.

🚀 Tier 1 (Free Models)

Available to all users with free registration. Free models have per-tier daily request limits — see Free Model Quota.

Manta Mini / Flash / Pro — MegaNova's dialogue-optimized model series with adaptive routing
GLM-4.7-Flash — efficient 30B MoE
DeepSeek-V3-0324-Economic — 685B economic model
L3-8B-Stheno-v3.2, L3-70B-Euryale-v2.1, L3.3-MS-Nevoria-70B, Sapphira-L3.3-70B-0.1, MN-Violet-Lotus-12B — community-tuned character dialogue models
Mistral-Small-3.2-24B-Instruct-2506 — 24B instruct model
Faster-Whisper-Large-V3 — speech-to-text
Qwen3-Embedding-8B — embeddings
BGE-Reranker-V2-M3 — reranking

Tier 1 Usage: Free models have daily limits and are intended for evaluation and interactive use, not high-volume production workloads. With "Allow paid usage after free quota" enabled, requests continue on a paid fallback once your daily quota is exhausted.

🎨 Tier 2+ (Premium Models)

Requires $1 deposit to unlock. Includes:

Multimodal: Claude 4 family (Opus / Sonnet / Haiku), GPT-5 family (5.4 / 5.3 / 5.2 / 5.1 / 5 / Mini / Nano / 4-Turbo / 4o-mini), Gemini 2.5 / 3 / 3.1, Grok-4-Fast, Qwen3.5/3.6-Plus, Kimi-K2.5, MiMo-V2-Omni, Mistral-Small-4, Gemma-4-31B
Text-only: GLM-5.1 / 5 / 4.7 / 4.6, MiMo-V2-Pro/Flash, MiniMax M2.1/2.5/2.7, DeepSeek V3.1/V3.2/V3-0324/R1-0528, Qwen3-235B, Kimi-K2-Thinking, Llama3.3-70B, L3.1/3.3-70B-Euryale, Cydonia-24B-v4.1, Mistral-Nemo-2407
Vision: Qwen3-VL-Plus/Flash, Qwen2.5-VL-32B
Image: Nano-Banana series, Bytedance Seedream 3.0–5.0-Lite, Alibaba Z-Image-Turbo, DreamShaper-8
Video: Google Veo-3.1 / Veo-3.1-Fast (T2V + I2V), Alibaba Wan 2.6 (T2V / I2V / I2V-Flash / R2V / R2V-Flash), BytePlus Seedance 1.0 Pro/Lite (T2V + I2V)
Audio: Gemini-2.5-Flash-TTS, Gemini-2.5-Pro-TTS

Pricing and Billing

Pay-As-You-Go: Charges are based on token usage (input + output) or per-image / per-second, depending on the model
Tier-Based Access: Paid models require Tier 2+; rate and daily request limits scale up with each tier
Free Models: Available to Tier 1 users with daily limits
Premium Models: Available to Tier 2+ users with higher RPD limits

Check the pricing page for detailed pricing information.