Replicate vs Together AI

Pricing, features, platforms and who each one suits — side by side.

Replicate

Run, fine-tune and deploy open-source ML models through a cloud API

PaidVisit site

Together AI

Cloud for running, fine-tuning and scaling open-source AI models on GPUs

PaidVisit site

Choose Replicate if…

you're in its core audience: software developers, ml engineers

Choose Together AI if…

you're in its core audience: ai startups, ml engineers

Side by side

FeatureR

ReplicateT

Together AI

Pricing modelPaidPaid

Starting price——

Free plan

Free trial

API

Open source

Data retentionRetained as long as necessary to provide Services; backups kept limited period—

MCP server—

Webhooks—

Official SDKs—

Self-hosting—

Custom code steps—

PlatformsWeb, APIWeb, API

CategoryDeveloper Tools & APIsLLM APIs & Model Hosting

CompanyReplicate · part of CloudflareTogether AI

Full feature lists

R Replicate

Thousands of public models for image, video, audio and text
Python, JavaScript and HTTP API clients
Fine-tune models such as SDXL on your own data
Deploy custom models packaged with open-source Cog
Autoscaling that scales to zero when idle
Per-second billing across CPU and Nvidia GPUs
Logging and monitoring of predictions

T Together AI

Serverless pay-per-token inference for open-source models
Batch inference up to 30 billion tokens per model
Provisioned throughput with 99% uptime SLA
Dedicated single-tenant model and container endpoints
On-demand and reserved NVIDIA H100, H200, B200, B300 clusters
LoRA and full fine-tuning with SFT and DPO
Code sandboxes and code interpreter for agents
Managed storage with zero egress fees

Their home pages

What each vendor shows first: the pitch, the plans and the sign-up.

Replicate home page — R Replicate · replicate.com

Together AI home page — T Together AI · together.ai

What you'd pay

The cheapest plan that fits, per month, billed yearly where the plan offers it.

Replicate

Pay-as-you-go (hardware)From $0.000025/sec (CPU Small, $0.09/hr); Nvidia T4 $0.81/hr; H100 $5.49/hr

Pay-as-you-go (per-model)e.g. flux-dev $0.025/image, flux-1.1-pro $0.04/image

EnterpriseCustom

Together AI

Serverless InferenceFrom $0.15 / $0.60 per 1M input/output tokens (gpt-oss-120B)

GPU Clusters (on-demand)$3.99/GPU-hour (H100)

Dedicated Inference$5.49/GPU-hour (H100)

Fine-tuningFrom $0.34 per 1M training tokens (Qwen3.5 9B SFT)

Provisioned ThroughputCustom

Price changes we've recorded

We re-read each pricing page every month and log what changed.

Replicate

No changes since we started tracking.

Together AI

Nov 8, 2026Serverless Inference: $0 → $0.15; New plan: GPU Clusters (on-demand) ($3.99); New plan: Dedicated Inference ($5.49); New plan: Fine-tuning ($0.34); Provisioned Throughput: Per PTU per minute → Custom; Plan dropped: Dedicated Inference (On-demand); Plan dropped: Dedicated Inference (Reserved)