Loading Comparison
Fetching pricing data and provider information...
Compare GPU and LLM inference API pricing between Microsoft Azure and Deep Infra. Find the best rates for AI training, inference, and ML workloads.
Provider 1
Provider 2
| GPU Model ↑ | Microsoft Azure Price | Deep Infra Price | Price Diff ↕ | Sources |
|---|---|---|---|---|
A100 SXM 80GB VRAM • Microsoft Azure | Not Available | — | ||
A100 SXM 80GB VRAM • | ||||
B200 180GB VRAM • Deep Infra | Not Available | — | ||
B200 180GB VRAM • | ||||
H100 NVL 94GB VRAM • Microsoft Azure | Not Available | — | ||
H100 NVL 94GB VRAM • | ||||
H100 SXM 80GB VRAM • Microsoft Azure | 8x GPU | Not Available | — | |
H100 SXM 80GB VRAM • $11.06/hour 8x GPU configuration Updated: 9/9/2026 ★Best Price Not Available | ||||
H200 141GB VRAM • Microsoft Azure | 8x GPU | Not Available | — | |
H200 141GB VRAM • $10.60/hour 8x GPU configuration Updated: 9/9/2026 ★Best Price Not Available | ||||
MI300X 192GB VRAM • Microsoft Azure | 8x GPU | Not Available | — | |
MI300X 192GB VRAM • $6.00/hour 8x GPU configuration Updated: 9/9/2026 ★Best Price Not Available | ||||
Tesla T4 16GB VRAM • Microsoft Azure | Not Available | — | ||
Tesla T4 16GB VRAM • | ||||
A100 SXM 80GB VRAM • Microsoft Azure | Not Available | — | ||
A100 SXM 80GB VRAM • | ||||
B200 180GB VRAM • Deep Infra | Not Available | — | ||
B200 180GB VRAM • | ||||
H100 NVL 94GB VRAM • Microsoft Azure | Not Available | — | ||
H100 NVL 94GB VRAM • | ||||
H100 SXM 80GB VRAM • Microsoft Azure | 8x GPU | Not Available | — | |
H100 SXM 80GB VRAM • $11.06/hour 8x GPU configuration Updated: 9/9/2026 ★Best Price Not Available | ||||
H200 141GB VRAM • Microsoft Azure | 8x GPU | Not Available | — | |
H200 141GB VRAM • $10.60/hour 8x GPU configuration Updated: 9/9/2026 ★Best Price Not Available | ||||
MI300X 192GB VRAM • Microsoft Azure | 8x GPU | Not Available | — | |
MI300X 192GB VRAM • $6.00/hour 8x GPU configuration Updated: 9/9/2026 ★Best Price Not Available | ||||
Tesla T4 16GB VRAM • Microsoft Azure | Not Available | — | ||
Tesla T4 16GB VRAM • | ||||
| Model ↑ | Microsoft Azure | Deep Infra | Input Diff ↕ |
|---|---|---|---|
Anthropic | Not available | $10.00 in $50.00 out | — |
Anthropic | Not available | $1.00 in $5.00 out | — |
Anthropic | Not available | $5.00 in $25.00 out | — |
Anthropic | Not available | $5.00 in $25.00 out | — |
Anthropic | Not available | $5.00 in $25.00 out | — |
Anthropic | Not available | $3.00 in $15.00 out | — |
Anthropic | Not available | $3.00 in $15.00 out | — |
DeepSeek | Not available | $0.500 in $2.15 out | — |
DeepSeek | Not available | $0.320 in $0.890 out | — |
DeepSeek | Not available | $0.250 in $0.950 out | — |
DeepSeek | Not available | $0.260 in $0.380 out | — |
DeepSeek | Not available | $0.090 in $0.180 out | — |
DeepSeek | Not available | $0.440 in $1.32 out | — |
DeepSeek | Not available | $1.30 in $2.60 out | — |
Google | Not available | $0.300 in $2.50 out | — |
Explore how these providers compare to other popular GPU cloud services
Compare Microsoft Azure with another leading provider
Compare Microsoft Azure with another leading provider
Compare Microsoft Azure with another leading provider
Compare Microsoft Azure with another leading provider
Compare Microsoft Azure with another leading provider
Compare Microsoft Azure with another leading provider
Unified AI platform with a catalog of models, agents, and tools for building and governing AI applications
Seamless integration with Microsoft ecosystem and enterprise tools
Strong hybrid and multi-cloud support with Azure Arc
Industry-leading security features and compliance certifications
Extensive worldwide network of data centers and edge locations
OpenAI-compatible API for 100+ models including DeepSeek, Qwen, Llama 4, Claude, and Gemini families with autoscaling
B200 instances with SSH access spin up in about 10 seconds and bill hourly
Deploy your own Hugging Face models onto dedicated A100, H100, H200, B200, or B300 GPUs
Published per-GPU hourly rates for A100, H100, H200, B200, and B300 with competitive pricing
Runs on its own inference-optimized hardware in US-based data centers rather than rented capacity
Support for text generation, vision and OCR, embeddings and reranking, image, video and music generation, and speech recognition and synthesis
GPU-enabled VMs for various workloads
Managed Kubernetes service with GPU support
End-to-end ML platform with GPU acceleration
Hosted model APIs with autoscaling on Deep Infra's own inference infrastructure.
On-demand GPU nodes with SSH access for custom workloads.
Multi-node B200 and B300 clusters with SSH access for training and full-control workloads.
Pay for compute capacity by the second, with no long-term commitments or upfront payments
Significant cost reduction versus pay-as-you-go rates in exchange for a one-year or three-year term; suited to stable, predictable workloads
Unused Azure compute capacity at deep discounts for interruptible workloads
Cost savings for existing Windows Server and SQL Server licenses
Save money across select compute services globally by committing to spend a fixed hourly amount for 1 or 3 years
OpenAI-compatible inference APIs billed per input and output token, with no idle GPU time or minimums
Published transparent hourly pricing for A100, H100, H200, B200, and B300 GPUs with pay-as-you-go billing
Non-LLM models are billed for inference execution time rather than per token
Prompt-cached input tokens are billed at a lower rate than uncached input tokens
Flexible hourly billing for dedicated instances with no prepayments, contracts, or minimums required
Sign up for Azure and get started with free credits
Configure your subscription, resource groups, and access controls
Select from VMs, containers, or serverless based on your needs
Launch your first GPU-enabled instance or AI service
Sign up (GitHub-supported) and open the Deep Infra dashboard
Add a payment method to unlock GPU rentals and API usage
Choose serverless APIs or dedicated A100, H100, H200, B200, or B300 instances
Start instances with SSH access or call the OpenAI-compatible API endpoints
Track spend and instance status from the dashboard and shut down when idle
60+ regions worldwide with multiple availability zones
Basic, Developer, Standard, and Professional Direct support plans with 24/7 options. Extensive documentation and community resources.
Runs on self-operated, inference-optimized infrastructure in US-based data centers; no international region list is published.
Documentation site, dashboard guidance, Discord community, feedback email (feedback@deepinfra.com), and contact-sales options.