EcoHash
RTX PRO 6000 GPU cloud and OpenAI-compatible inference API
Last reviewed Aug 28, 2026
EcoHash rents NVIDIA RTX PRO 6000 Server Edition (Blackwell, 96 GB) GPUs as instances of 1, 2, 4 or 8 GPUs and multi-replica clusters, and serves open models for text, vision, image, speech and video through an OpenAI-compatible inference API. It also offers dedicated inference endpoints and managed LoRA/QLoRA fine-tuning. EcoHash Platform Technology LLC was founded in 2025 as the HPC and AI inference division of Cango Inc. (NYSE: CANG), a Bitcoin mining company.
Available GPUs
Hourly on-demand pricing. Click column headers to sort.
Prices last updated: October 12, 2026
| RTX PRO 6000 | 1×2×4×8× | 1×2×4×8× |
EcoHash pricing by GPU
Configurations, price rank, and alternatives for one GPU at a time.
LLM API Pricing
Pay-per-token pricing. Prices shown per 1M tokens.
Prices last updated: October 12, 2026
| Model | Input/1M | Output/1M | ||||
|---|---|---|---|---|---|---|
| $0.050 | $0.200 | |||||
| $0.090 | $0.370 | |||||
| $0.100 | $0.300 | |||||
| $0.100 | $0.100 | |||||
| $0.150 | $0.500 | |||||
| $0.200 | $0.900 | |||||
| $0.400 | $0.800 | |||||
| $0.400 | $0.400 | |||||
| $0.490 | $1.42 | |||||
| $0.910 | $2.72 | |||||
EcoHash pricing by model
Input, output, and batch rates, plus alternatives, for one model at a time.
Pros & Cons
Advantages
- Per-second billing with a single per-GPU rate across instance sizes
- 96 GB of VRAM per GPU on Blackwell-generation RTX PRO 6000 cards
- One OpenAI-compatible API for platform models and dedicated endpoints
- Covers GPU rental, fine-tuning and inference on one account
- Public per-token and per-GPU-hour pricing
Limitations
- Single GPU model (RTX PRO 6000); no H100, H200 or B200 capacity
- GPU instances available in the United States only
- Fine-tuning limited to LoRA/QLoRA jobs on up to 4 GPUs
- Founded in 2025, so a short operating history
Key Features
RTX PRO 6000 GPU Instances
Containerized Linux instances with 1, 2, 4 or 8 RTX PRO 6000 GPUs, with root access, SSH, a browser terminal and Jupyter.
GPU Clusters
Clusters of 2–8 identical replicas launched as one unit.
Per-Second Billing
GPU instances are billed per second of running time at one per-GPU rate across instance, dedicated-inference and cluster workloads.
OpenAI-Compatible Inference API
Pay-per-token access to platform models for chat, embeddings, reranking, image generation, speech-to-text, text-to-speech, voice cloning and video generation at api.ecohash.com.
Dedicated Inference
Deploy a Hugging Face or registered model as a managed endpoint called through the same API and key.
Managed Fine-Tuning
LoRA/QLoRA fine-tuning jobs on 1, 2 or 4 GPUs, billed per minute, with a dataset registry that validates JSONL before any GPU is booked.
Storage
Attachable cloud drives and shared filesystems for instances.
Multi-Region Routing
Inference requests are routed across regions with DNS failover for dedicated endpoints.
Compute Services
RTX PRO 6000 GPU Instances
Containerized Linux GPU instances with SSH, web-terminal and Jupyter access, billed per second at the same per-GPU rate for every size.
Pricing Options
| Option | Details |
|---|---|
| GPU instances | Per GPU-hour, billed per second of running time |
| Inference API | Per token for language models and per second for some media models, with no minimum commitment |
| Dedicated Inference | Per GPU-hour for the GPUs behind the endpoint |
| Fine-tuning | Per minute of GPU time |
Availability & Support
Regions
GPU instances: United States. Inference API: multi-region routing.
Support
Documentation at docs.ecohash.com and a contact form on ecohash.com; an SLA is published on the website.
Company
Published by EcoHash to identify the company behind this listing.
- Legal entity
- EcoHash Platform Technology LLC
Getting Started
- 1
Create an account
Sign up at console.ecohash.com.
- 2
Create an API key
Generate a key in the console for the inference API and dedicated endpoints.
- 3
Call a platform model
Point an OpenAI SDK client at https://api.ecohash.com/v1 and choose a model from the catalog.
- 4
Launch a GPU instance
Pick a compute template and a 1, 2, 4 or 8 GPU size, then connect over SSH, the browser terminal or Jupyter.
Frequently Asked Questions
What GPU types does EcoHash offer?
EcoHash offers various GPU types including RTX PRO 6000, RTX PRO 6000, RTX PRO 6000, RTX PRO 6000. Check the pricing table above for current availability and pricing.
How do I get started with EcoHash?
Create an account, Create an API key, Call a platform model, Launch a GPU instance
What are EcoHash's main advantages?
EcoHash's main advantages include: Per-second billing with a single per-GPU rate across instance sizes, 96 GB of VRAM per GPU on Blackwell-generation RTX PRO 6000 cards, One OpenAI-compatible API for platform models and dedicated endpoints, Covers GPU rental, fine-tuning and inference on one account, Public per-token and per-GPU-hour pricing.
What are EcoHash's limitations?
EcoHash's main limitations include: Single GPU model (RTX PRO 6000); no H100, H200 or B200 capacity, GPU instances available in the United States only, Fine-tuning limited to LoRA/QLoRA jobs on up to 4 GPUs, Founded in 2025, so a short operating history.
Compare Providers
Find the best prices for the same GPUs and models from other providers
