# ComputeUnion

> ComputeUnion is a real-time GPU rental and AI API price comparison platform covering 70+ providers and 3,400+ AI model APIs, updated automatically every 6 hours.

## What it does
- Compare cloud GPU rental prices (H100, A100, RTX 4090, L40S, A6000, etc.) — on-demand and spot pricing in USD/hr across 20+ platforms
- Compare AI API token prices for LLM models (GPT, Claude, Gemini, DeepSeek, Llama, Qwen and 100+ more) — input/output per 1M tokens
- Compare video generation API pricing (Kling, VEO 3, Seedance, Runway, Pika, Luma, Hailuo, Wan)
- Compare audio/speech API pricing — TTS and ASR (OpenAI, ElevenLabs, MiniMax, Deepgram, Cartesia)
- Compare image generation API pricing (GPT Image 1, DALL-E 3, FLUX 1.1 Pro, Midjourney v7, Ideogram, Stability AI)
- Track upcoming AI model releases and pre-publish pricing pages before launch
- Track active promotions, free credits, and discounts from GPU and AI API platforms
- GPU Advisor: AI-powered recommendation tool matching workloads to the optimal GPU and platform

## Key Pages

### English
- [Home](https://www.computeunion.net): Real-time GPU price leaderboard, cost estimator, platform comparison
- [LLM API Pricing](https://www.computeunion.net/api-pricing): 70+ providers, sorted cheapest first — input/output per 1M tokens
- [Video Gen API Pricing](https://www.computeunion.net/api-pricing/video): Kling, VEO 3, Seedance, Runway, Pika, Luma, Hailuo — per second or per video
- [Audio API Pricing](https://www.computeunion.net/api-pricing/audio): TTS, ASR, realtime speech models
- [Image Gen API Pricing](https://www.computeunion.net/api-pricing/image): GPT Image 1, DALL-E 3, FLUX, Midjourney — per image
- [Upcoming Models](https://www.computeunion.net/api-pricing/coming-soon): Pre-launch pages for announced but not-yet-released models
- [GPU Rental Prices](https://www.computeunion.net/gpu): All GPU models with cross-platform rental price comparison
- [Platforms](https://www.computeunion.net/platform): All 70+ GPU and API platforms with pricing overview
- [Promotions](https://www.computeunion.net/activities): Active GPU and AI API credits and discounts
- [GPU Advisor](https://www.computeunion.net/advisor): Describe your workload, get GPU + platform recommendations

### Chinese (中文)
- [首页 GPU算力价格](https://www.computeunion.net/zh): 实时GPU云算力价格排行榜
- [语言模型API价格](https://www.computeunion.net/zh/api-pricing): GPT、Claude、Gemini、DeepSeek等，70+服务商，百万Token费用对比，含国内中转站
- [视频生成API价格](https://www.computeunion.net/zh/api-pricing/video): 可灵、VEO 3、Seedance、Runway等视频API价格
- [语音API价格](https://www.computeunion.net/zh/api-pricing/audio): TTS、ASR语音模型价格对比
- [图像生成API价格](https://www.computeunion.net/zh/api-pricing/image): DALL-E、FLUX、Midjourney等图像生成API价格
- [即将发布模型](https://www.computeunion.net/zh/api-pricing/coming-soon): 已公告但尚未发布的AI模型预览页
- [GPU算力租赁价格](https://www.computeunion.net/zh/gpu): 各GPU型号跨平台租赁价格对比
- [GPU平台](https://www.computeunion.net/zh/platform): 70+个GPU云算力与AI API平台列表

## GPU Platforms
Vast.ai, RunPod, Lambda Labs, Paperspace, TensorDock, FluidStack, Vultr, Modal Labs, Latitude.sh, DataCrunch, Jarvis Labs, Oblivus, Cudo Compute, Nebius AI, Hyperbolic, Cerebras, SambaNova

## AI API Providers

### Official APIs
OpenAI (GPT-4o, GPT-4.1, o3, o4-mini), Anthropic (Claude 4, Claude Sonnet, Claude Haiku), Google (Gemini 2.5 Pro, Gemini 2.5 Flash, Vertex AI), DeepSeek (V3, R1), xAI (Grok), Mistral, Cohere, Perplexity, AI21 Labs, Reka AI, Writer (Palmyra)

### Global Aggregators
OpenRouter, Groq, Fireworks AI, Together AI, Lepton AI, Novita AI, Cerebras, DeepInfra, Hyperbolic, Cloudflare AI, Nebius AI, Chutes.ai

### Chinese Official APIs
SiliconFlow (硅基流动), Zhipu AI (智谱/GLM), Moonshot (Kimi), MiniMax, Baichuan AI (百川), Doubao (豆包/ByteDance), Hunyuan (腾讯混元), iFlytek Spark (讯飞星火), StepFun (阶跃星辰), Yi (零一万物), DashScope (阿里云百炼), Qianfan (百度千帆)

### Chinese Relay Stations (国内中转站)
302.AI, AiHubMix, YunWu AI (云雾AI), API2D, OhMyGPT, EasyRouter, IKunCode, dddAI, PackyCode, RunAPI, ProAI API, PoloAPI

## Update Schedule
- GPU spot prices: every 6 hours
- GPU on-demand prices: every 24 hours
- AI API prices: every 6 hours
- Upcoming model pages: within 24 hours of official announcement

## Data API
- GPU leaderboard: https://www.computeunion.net/api/gpu-leaderboard
- Platform activities: https://www.computeunion.net/api/activities