Inference providers, head to head.

What each provider does, what the other does, and which one to pick for your workload.

Compare by provider

Pick a provider to see how it stacks up against every other one.

OpenAIThe most widely adopted closed-model API, from GPT-6 Astra down to GPT-6 Luna.

AnthropicClosed Claude models known for agentic coding, with 1M context across the lineup.

Google Vertex AIGoogle Cloud's enterprise AI platform, with Gemini and Claude next to a full MLOps stack.

Amazon BedrockAWS's managed model service: Claude, GPT and open models inside your AWS security posture.

Together AIThe broadest open-model platform: serverless, dedicated, GPU clusters and fine-tuning on one bill.

Fireworks AIFast open-model inference plus SFT, DPO and RL training at base-model prices.

BasetenLowest measured time to first token, with OpenAI and Anthropic compatible endpoints.

GroqExtreme speed on its own LPU chip, with tight tail latency on a small catalog.

CerebrasThe fastest public inference host, running models on a wafer-scale chip.

DeepInfraThe price floor for open-model inference, with 200+ models and no minimums.

Hugging Face Inference ProvidersOne Hugging Face token and one OpenAI compatible router in front of 17 partner inference providers, at provider rates.

ModalServerless GPU compute for Python: bring your own model and pay by the second.

Cloudflare Workers AIServerless open-model inference on Cloudflare GPUs, called straight from Workers and billed in Neurons.

xAIClosed Grok models with native live data from X and cheap output tokens.

Mistral AIEuropean lab shipping open-weight models on its own API, every major cloud, or your GPUs.

DeepSeekMIT-licensed open models at some of the lowest first-party prices anywhere.

Moonshot AIThe lab behind Kimi K3 and K3.1 Pro, among the most capable open-weight models, with 1M context.

Z.aiMaker of the open-weight GLM models, with a cheap coding plan.

Alibaba CloudHome of the Qwen family, with an open-weight Max flagship inside a full public cloud.

MetaThe Muse model family on a new closed API, plus open-weight Muse Spark Open and Glimmer.

CohereEnterprise-first models for RAG, search and agents, built to run in a private cloud or on-prem.

SambaNovaFast decode on large open models, served on its own dataflow chip.

NebiusA European AI cloud with managed open-model inference and raw GPUs on one account.

CrusoeAn energy-first AI cloud that builds its own data centers and serves open models on a cluster-wide KV cache.

falThe go-to platform for generative media, with 1,000+ image, video and audio models.

Novita AIA low-cost inference cloud with 200+ open models across text, image, video and speech.

VeniceA privacy-first AI platform with an OpenAI compatible API, uncensored open models and token-staked API credits.

ParasailAggregated GPUs behind one API, with cheap batch on any Hugging Face model.

Inference.netCheap batch inference on spare GPU capacity, plus a path from traces to custom models.

GMI CloudA GPU cloud that owns its hardware, with APAC data residency and 200+ models.

Thinking MachinesTinker, a post-training API for open-weight models, plus the open Inkling models.

Sail ResearchCheap open-model inference: pick a completion window and pay 30 to 85% less.

MorphSmall specialist models that apply coding-agent edits at 10,500+ tokens per second.

RelaceLow-cost open-model inference, plus fast small models for coding agents.

TypeSafe AIDecision models that return typed answers with calibrated confidence in milliseconds.

StepFunA Shanghai lab shipping efficient multimodal models, many under Apache 2.0.

RunwareLow-cost media generation across image, video, audio and 3D behind one request schema.

StreamLakeKuaishou's AI cloud, home of the KAT-Coder agentic coding models.

WaferAgent-tuned inference stacks that run open models faster on the same weights.

RunInfraOpen models for agents, plus an agent that benchmarks and builds deployments for you.

Particle.AIAn early startup serving cheap Flash-class open models with 1M context.

LuminalAn open-source compiler that turns models into native GPU code, sold as serverless or on-prem.

InfronAn AI gateway with 400+ models behind one API, at provider rates with failover and region pinning.

InceptionThe only commercial diffusion LLMs, generating tokens in parallel for 1,000+ tok/s on standard GPUs.

DolphinUncensored open models and a decentralized network that serves them on idle consumer GPUs.

OpenScaleA Y Combinator-backed API selling DeepSeek V4.1 Flash at some of the lowest prices around.

General ComputeAn inference cloud that runs decode on Cerebras, SambaNova and other ASICs instead of GPUs.

BoundlessAn inference partner that serves open models per token and tunes reserved capacity to your traffic.

How we compare providers

Profiles draw on each provider’s own docs and pricing pages plus third-party reviews and benchmarks, linked as sources on every profile. Subconscious is one of the providers compared, so each head-to-head also says where the other side wins. Pricing and model lineups change often; treat figures as a snapshot from the date above.

By The Subconscious Team · Updated October 29, 2026

Inference providers, head to head.

Subconscious vs the field

Compare by provider

How we compare providers