Inference APIs
OpenAI-compatible · chat, speech, transcription

Run popular AI models with one simple API.

Chat, text-to-speech and transcription on one OpenAI-compatible endpoint — GPT-OSS 120B, DeepSeek V4 Flash, Kokoro, Whisper and more. Pay per request, no tiers, no daily caps.

Fast to integrate

Familiar chat completions format. Send a request with cURL, Python or JavaScript in minutes.

Pay as you go

Prepaid credits with no subscription or minimums.

Zero data retention

US-hosted. Prompts and outputs are not kept by us or by the GPU providers, and nothing trains a model. How models are served.

Models

View all
GPT-OSS 120B
Text → Text · 131K context
Available

OpenAI's open-weight 120B reasoning model. Strong at coding, tool use and agentic tasks; the drop-in replacement Groq recommends for Llama 3.3 70B.

DeepSeek V4 Flash
Text → Text · 1M context
Available

DeepSeek's fast, inexpensive V4 model with a 1M-token context. Strong general assistant and coding model at the lowest price per token here.

DeepSeek V4.1 Flash
Text → Text · 1M context
Available

The newest DeepSeek Flash model: much quicker to answer than V4 Flash in our measurements, with a 1M-token context.

GLM 5.3 Flash
Text → Text · 1M context
Available

Z.ai's fast GLM 5.3 model with tool calling, JSON mode and a 1M-token context, at a low per-token price.

Llama 3.3 70B Instruct
Text → Text · 131K context
Available

Meta's instruction-tuned 70B model. Still served here after its retirement on Groq — the same llama-3.3-70b-versatile id works unchanged.

GLM 5.3
Text → Text · 1M context
Available

Z.ai's full-size GLM 5.3 model: tool calling, JSON-schema output and a 1M-token context, for coding agents and long tasks.

DeepSeek V4 Pro
Text → Text · 1M context
Available

DeepSeek's largest V4 model (0813 release) with a 1M-token context, hosted in the United States.

MiniMax M3
Text → Text · 512K context
Available

MiniMax's M3 model: tool calling and a 512K-token context at a price close to the Flash tier.

Kimi K2.7 Code
Text → Text · 262K context
Available

Moonshot AI's coding-tuned Kimi model with a 262K-token context, for coding agents and IDE assistants.

GPT-OSS 20B
Text → Text · 131K context
Available

OpenAI's small open-weight reasoning model: the cheapest chat model here, for lightweight assistants, classification and extraction.

Qwen3.8 Flash
Text → Text · 1M context
Available

Alibaba's fast Qwen3.8 model with a one-million-token context window and cached-input pricing.

Qwen3-VL 235B
Image + Text → Text · 262K context
Available

Alibaba's large vision-language model: send images in a chat request for OCR, document extraction and screenshot reasoning.

Kimi K3.1 Pro
Text → Text · 1M context
Available

Moonshot AI's flagship Kimi model with a 1,048,576-token window, served with the thinking switch and tool calling verified.

Kimi K2.6
Text → Text · 262K context
Available

Moonshot AI's Kimi K2.6 agentic model, served for the many projects that pin this exact id, with a 262K-token window.

MiniMax M3.1 Flash
Text → Text · 1M context
Available

MiniMax's M3.1 Flash: a 1,048,576-token reasoning model that streamed at 146 tokens per second here, priced with the Flash tier.

Nemotron 3.5 Super VL
Image + Text → Text · 262K context
Available

NVIDIA's open vision-language model: images in the OpenAI image_url format, tool calling and JSON output, at a Flash-tier price.

Kokoro 82M
Text → Audio
Available

Fast, natural open-weight text-to-speech with 54 voices across 8 languages, on the OpenAI /v1/audio/speech endpoint.

Orpheus 3B
Text → Audio
Available

Expressive Llama-based text-to-speech with emotive tags and eight English voices.

Qwen3 TTS
Text → Audio
Available

Alibaba's Qwen3 text-to-speech: nine preset voices, ten languages and instruction-controlled delivery, on the OpenAI /v1/audio/speech endpoint.

Whisper Large v3
Audio → Text
Available

OpenAI's multilingual speech recognition model on the standard /v1/audio/transcriptions endpoint — same request shape as OpenAI and Groq.

Parakeet TDT 0.6B v3
Audio → Text
Available

NVIDIA's very fast English-first transcription model with accurate timestamps; ideal for long recordings and batch jobs.

Voxtral Small 24B
Audio → Text
Available

Mistral's 24B speech model as a transcription endpoint: eight languages with automatic detection, strong on accented and noisy audio.

Voxtral Mini 3B
Audio → Text
Available

Mistral's small Voxtral model: the lowest price per minute here, eight languages with automatic detection.

BGE-M3
Text → Vector · 8K context
Available

BAAI's multilingual embedding model: 1,024-dimension vectors, 8K-token inputs, on the OpenAI /v1/embeddings endpoint.

Qwen3 Embedding 8B
Text → Vector · 32K context
Available

Qwen's 8B embedding model: up to 4,096 dimensions with adjustable size, 32K-token inputs, multilingual.

Qwen3 Reranker 8B
Query + Texts → Scores · 32K context
Available

Qwen's 8B reranker: scores passages against a query so RAG pipelines keep the right context.

OmniParser V2
Image → Structured data
Unavailable

Parse screenshots into structured, labeled UI elements for computer-use agents.

Quickstart

Create an account and get your API key

Sign up; your key is ready immediately on the API Keys page, and you add funds when you are ready to call it.

Make your first request
curl https://api.inferenceapis.com/v1/chat/completions \
  -H "Authorization: Bearer $INFERENCE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-oss-120b",
    "messages": [
      {"role": "system", "content": "You are a helpful assistant."},
      {"role": "user", "content": "Hello! What can you do?"}
    ]
  }'
import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.inferenceapis.com/v1",
    api_key=os.environ["INFERENCE_API_KEY"],
)
response = client.chat.completions.create(
    model="openai/gpt-oss-120b",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Hello! What can you do?"},
    ],
)
print(response.choices[0].message.content)
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.inferenceapis.com/v1",
  apiKey: process.env.INFERENCE_API_KEY,
});
const response = await client.chat.completions.create({
  model: "openai/gpt-oss-120b",
  messages: [
    { role: "system", content: "You are a helpful assistant." },
    { role: "user", content: "Hello! What can you do?" },
  ],
});
console.log(response.choices[0].message.content);
Explore

Tune prompts in the playground, browse the model catalog, or read the API reference.