43 Inference & Compute AI tools
AI services in the Inference & Compute field. Click a name to see its description, pricing and alternatives.
Inference & Compute: free vs paid at a glance
18 tools with official pricing data. 9 have a free plan, 9 offer a Korean UI. Cheapest paid plan: LM Studio ($20/mo).
| Tool | Free | Cheapest monthly | Korean | Where |
|---|---|---|---|---|
| AWS Bedrock | No | — | Yes | Web, API, CLI |
| AWS SageMaker | No | — | Yes | Web, API, CLI |
| Azure ML | No | — | Yes | Web, API |
| Cerebras WSE | Yes | — | No | Web, API |
| Cloudflare AI | Yes | — | Yes | Web, API |
| CUDA | Yes | — | No | Windows |
| Databricks | Yes | — | Yes | — |
| Fireworks AI | No | — | No | Web, CLI, API |
| Google Vertex AI | Yes | — | Yes | Web, API |
| Heurist | Yes | — | No | Web, API |
| LM Studio | Yes | $20/mo | No | CLI, API |
| Nebius | No | — | No | Web, CLI |
| ollama | Yes | $20/mo | — | Windows, Mac, CLI |
| RunPod | No | — | No | API |
| Seldon Core | No | — | No | API |
| SillyTavern | Yes | — | Yes | Windows, Mac, Linux |
| TensorRT | No | — | Yes | Windows, API |
| Triton | No | — | Yes | Windows, API |
From each tool's official pricing page (checked by 2026-10-14). '—' means not stated. Prices change, so check before paying.
How to choose
- Try a tool with a free plan first and check the limit before paying.
- If you need Korean menus, filter by the Korean column.
- Annual plans are often cheaper; the table shows monthly prices.
Akash Network
Decentralized cloud compute marketplace for AI workloads on GPU nodes
Read the guide →FreeApple MLX
Array framework for Apple Silicon ML research
Read the guide →Free + paidAWS Bedrock
AWS managed service for foundation models from Anthropic, Meta, and Mistral
Read the guide →Free + paidAWS SageMaker
Amazon's fully managed ML platform
Read the guide →Free + paidAzure ML
Microsoft Azure's cloud ML service
Read the guide →Free + paidBentoML
Framework for building ML-powered services
Read the guide →Free + paidCerebras WSE
Wafer-scale engine for fastest LLM training and inference
Read the guide →Free + paidCloudflare AI
AI inference at the edge via Cloudflare Workers AI with zero cold starts
Read the guide →Free + paidCoreWeave
GPU-accelerated cloud for AI/ML workloads
Read the guide →Free + paidCUDA
NVIDIA's parallel computing platform
Read the guide →Free + paidDatabricks
Unified data analytics and AI platform
Read the guide →Free + paidDeepSpeed
Deep learning optimization library by Microsoft
Read the guide →FreeEXO
Run frontier AI on your own devices by clustering everyday hardware
Read the guide →Free + paidFastChat
Distributed multi-model LLM serving system with OpenAI-compatible API and web UI
Read the guide →Free + paidFireworks AI
Fast, affordable inference API for open-source models with sub-100ms latency
Read the guide →Free + paidGoogle Vertex AI
Google Cloud's unified ML platform
Read the guide →Free + paidGroq LPU
Language Processing Unit for ultra-fast LLM inference
Read the guide →Free + paidHeurist
Decentralized GPU network for low-latency, verifiable AI model inference
Read the guide →Free + paidHorovod
Distributed deep learning training framework
Read the guide →Free + paidllama.cpp
LLM inference in C/C++ for local deployment
Read the guide →Free + paidLM Studio
Desktop app for discovering, downloading, and running open-weight LLMs locally
Read the guide →Free + paidMegatron-LM
NVIDIA's large-scale transformer training
Read the guide →FreeNanotron
Minimalist large language model training by Hugging Face
Read the guide →Free + paidNebius
AI cloud with GPU clusters and Token Factory managed inference
Read the guide →Free + paidollama
Run local LLMs (Llama, Qwen, Gemma, and more) with a single command
Read the guide →Free + paidONNX Runtime
Cross-platform ML model accelerator
Read the guide →Free + paidOpen WebUI
Self-hosted chat UI for local and cloud LLMs; Ollama, OpenAI-compatible API, RAG
Read the guide →Free + paidROCm
AMD's open software platform for GPU computing
Read the guide →Free + paidRunPod
Developer-focused GPU cloud for training, inference, and serverless AI
Read the guide →Free + paidSambaNova
Full-stack AI platform with custom RDU chips for fast inference at scale
Read the guide →Free + paidSeldon Core
Kubernetes-native ML deployment platform
Read the guide →Free + paidSGLang
Fast serving framework for LLMs and VLMs with structured generation support
Read the guide →Free + paidSillyTavern
Locally installed UI for text, image, and voice LLMs with deep customization
Read the guide →Free + paidSol Engine
NVIDIA's training-free acceleration framework for video diffusion inference
Read the guide →Free + paidTensorRT
NVIDIA's SDK for high-performance inference
Read the guide →Free + paidTensorRT-LLM
NVIDIA's LLM inference framework for NVIDIA GPUs with TensorRT acceleration
Read the guide →Free + paidTenstorrent
AI hardware with open-source software stack
Read the guide →Free + paidTorchServe
PyTorch model serving framework
Read the guide →Free + paidTriton
NVIDIA's open-source inference serving software
Read the guide →Free + paidvLLM
High-throughput and memory-efficient LLM serving
Read the guide →Free + paidFSDP
Fully Sharded Data Parallel training in PyTorch
View on the 3D map →Free + paidHarbor
Effortlessly run LLM backends, APIs, frontends, and services with one command
View on the 3D map →Free + paidRay Serve
Scalable model serving with Ray
View on the 3D map →