Geodd
Inference API, dedicated endpoints and bare-metal GPUs
Geodd is a US AI inference company that serves open-weight text, image and video models through an OpenAI-compatible serverless API, and also offers dedicated inference endpoints and bare-metal GPU rental. It develops hardware-specific kernels with its own kernel-development models and deploys the improvements that pass verification into its serving stack.
At a glance · prices collected Nov 19, 2026
What's good about Geodd
- OpenAI-compatible API with a separate EU endpoint
- Inference inputs and outputs are processed in memory and not stored or used for training, per its data privacy documentation
- Serverless inference, dedicated endpoints and bare-metal GPUs under one account
- Model catalogue with per-token rates published through an unauthenticated API
- 24/7 support with a stated target first response of 10 minutes
Geodd LLM API prices
Pay-per-token pricing, per 1M tokens.
Prices last updated November 19, 2026
| Model | Input/1M | Output/1M |
|---|---|---|
| $0.039 | $0.158 | |
| $0.110 | $0.400 | |
| $0.130 | $0.370 | |
| $0.140 | $0.300 | |
| $0.200 | $2.50 | |
| $0.300 | $1.10 | |
| $0.600 | $3.20 | |
| $0.900 | $2.80 | |
| $1.00 | $4.30 | |
| $1.48 | $3.40 |
Geodd pricing by model
Input, output, and batch rates, plus alternatives, for one model at a time.
Billing & terms
- Serverless per-token pricing
- Text models billed per 1M input and output tokens; image models billed per image
- On-demand dedicated GPUs
- Published hourly rates for single- and multi-GPU configurations
- Reserved capacity
- Reservation-only GPUs with pricing on request
- Prepaid wallet or invoiced billing
- Standard accounts fund a prepaid wallet with optional auto top-up; Enterprise accounts are invoiced
- Startup credits
- Credits for accelerator-backed and early-stage teams, usable across inference and dedicated GPUs for 12 months
- Regions
- Serverless inference in US East and Norway, with a dedicated EU endpoint; GPU VM regions us-east-1 (US) and lk-west-1 (APAC)
- Support
- 24/7 email support with direct engineer access and a target first response of 10 minutes (not a guaranteed SLA unless contracted); documentation, changelog and status page
Watch out for
- The GPU API is Enterprise-only, and instance management through it is not yet available
- Standard accounts are prepaid through a wallet; invoiced billing requires an Enterprise account
- B200, B300 and MI300X are reservation-only with multi-week lead times
- SOC 2 is listed as pending
Features
- Serverless inference API
- OpenAI-compatible endpoints for chat completion, completion, embedding and image generation, billed per token for text models
- US and EU base URLs
- A global (US) base URL and a separate EU base URL for regional data processing
- Dedicated inference
- Single-tenant model endpoints on dedicated GPU allocation for sustained-throughput workloads
- Dedicated GPU instances
- Bare-metal GPU servers provisioned from the Deploy Pad console with SSH access and full OS control; B200, B300 and MI300X are offered by reservation
- Kernel optimization
- AI agents using hardware-specific models (Meridian for NVIDIA CUDA; Helix for AMD ROCm and Stride for Tenstorrent announced) write and test kernels against production workloads before release
- Optimized local runtimes
- Free model-specific inference runtimes with Geodd-developed kernels for running on your own hardware
- CLI and MCP access
- Documented CLI and MCP guides alongside the HTTP API
Company
Published by Geodd to identify the company behind this listing.
- Legal entity
- Geodd LLC
Getting started
Step 1
Create an account
Sign up and open the Deploy Pad console
Step 2
Fund the wallet
Add funds under Billing > Manage Funds, or set up auto top-up with a saved card
Step 3
Create an API key
In Deploy Pad, go to Serverless Inference and press Create API Key
Step 4
Send a request
Point an OpenAI-compatible client at the global or EU base URL and pick a model from the Available Models list
Where this page's facts come from
- LLM API prices
- geodd.io · last collected Nov 19, 2026
- Profile
- Reviewed Aug 20, 2026
- Method
- How we collect and check prices
Frequently Asked Questions
Which models does Geodd serve?
Check the pricing table above for the models Geodd currently serves and their per-token rates.
How do I get started with Geodd?
Create an account, Fund the wallet, Create an API key, Send a request
What are Geodd's main advantages?
Geodd's main advantages include: OpenAI-compatible API with a separate EU endpoint, Inference inputs and outputs are processed in memory and not stored or used for training, per its data privacy documentation, Serverless inference, dedicated endpoints and bare-metal GPUs under one account, Model catalogue with per-token rates published through an unauthenticated API, 24/7 support with a stated target first response of 10 minutes.
What are Geodd's limitations?
Geodd's main limitations include: The GPU API is Enterprise-only, and instance management through it is not yet available, Standard accounts are prepaid through a wallet; invoiced billing requires an Enterprise account, B200, B300 and MI300X are reservation-only with multi-week lead times, SOC 2 is listed as pending.
Compare Providers
Find the best prices for the same GPUs and models from other providers · All Geodd alternatives, ranked →
All Geodd Comparisons
The weekly price report
One email every Monday with the week's notable GPU and LLM API price moves, plus occasional product news. No spam.