What Open-Source Models Does Together AI Support?

Together AI hosts 200+ open-source models including DeepSeek V3 and R1, Llama 3.3 and 4, Qwen3 series, Gemma 4, FLUX image models, Kling video models, and speech models like Whisper and Cartesia Sonic. New models are added regularly as they become available.

Is Together AI API Compatible With OpenAI SDK?

Yes, Together AI provides an OpenAI-compatible API, meaning developers can switch from OpenAI to Together AI by changing the base URL and API key. Existing code using the OpenAI Python or TypeScript SDK works with minimal modification.

How Does Together AI Fine-Tuning Work?

Together AI supports supervised fine-tuning (SFT) and direct preference optimization (DPO) with both LoRA and full fine-tuning options. Pricing is per million tokens processed. Supported models include DeepSeek, Llama, Qwen, and GLM families, with no training infrastructure to manage.

What GPU Hardware Is Available For Dedicated Endpoints?

Together AI offers NVIDIA H100 80GB, H200 141GB, and B200 180GB GPUs for dedicated endpoints. On-demand pricing starts at $3.99/hr for H100 and $9.95/hr for B200. Reserved commitments of 7-180+ days offer discounted rates.

Does Together AI Store User Prompts Or Training Data?

Together AI offers a Zero Data Retention (ZDR) mode that can be enabled in privacy settings. When active, prompts, model responses, and submitted content are not stored or used for model training. Users retain full control over data handling preferences.

Together AI

Name: Together AI
Availability: InStock
Author: Together Computer, Inc.

by Together Computer, Inc. • San Francisco, CA, USA • Founded 2022

Full-Stack AI Native Cloud for Open-Source Model Inference

No reviews yet

14 1

ai-inference open-source-models gpu-cloud model-fine-tuning llm-api

Visit Together AI

Follow:

Trust Score

Based on ratings & reviews

3.5 /5

38 reviews

Views

Clicks

Pricing

Pay-as-you-go

What is Together AI?

Together AI is a full-stack cloud platform built for running, fine-tuning, and training open-source AI models at production scale. It offers serverless inference across 200+ models — including DeepSeek, Llama, Qwen, and Gemma families — with an OpenAI-compatible API.

The platform provides dedicated GPU endpoints on NVIDIA H100, H200, and B200 hardware, async Batch API processing for up to 30 billion tokens, and managed fine-tuning with LoRA and DPO support. Backed by foundational research like FlashAttention and speculative decoding, Together AI delivers 2x faster inference and up to 60% lower costs compared to standard cloud providers.

Together AI — Full-Stack AI Native Cloud for Open-Source Model Inference Whether you're evaluating Together AI for your team or comparing it to alternatives in the AI Business Tools category, this in-depth review covers everything: features, pricing, real user reviews, pros and cons, integrations, and direct comparisons against competitors.

Together AI Demo Video

Key Features 8

2x Faster Open-Source Model Inference Powered by Cutting-Edge Research

Serverless, Batch, and Dedicated Inference for Scalable AI Deployment

200+ AI Models for Text, Image, Video, Code, and Audio

GPU Clusters With NVIDIA H100, H200, B200, and GB200 Support

Fine-Tuning With SFT, DPO, LoRA, and Reinforcement Learning API

FlashAttention-4 Kernel Delivering 4x Long-Context Performance Improvement

Secure Code Sandboxes for AI Agents and App Development

Up to 60% Lower Cost With Workload-Specific Optimization and Batch API

Who Is Together AI For

1 AI Engineers

2 ML Researchers

3 Backend Developers

4 AI Startup Teams

5 DevOps Engineers

6 Data Scientists

Pros & Cons

Pros 4 benefits

Research-Backed Inference Optimization
Extensive Open-Source Model Catalog
Flexible Pay-Per-Token Pricing
SOC 2 Type 2 Certified

Cons 3 limitations

Requires Developer Technical Expertise
No Native GUI Playground
Complex Pricing Structure Navigation

Frequently Asked Questions

5 questions

How Together AI works

Together AI is positioned as full-Stack AI Native Cloud for Open-Source Model Inference. Under the hood it ships 8 headline capabilities, including 2x Faster Open-Source Model Inference Powered by Cutting-Edge Research, Serverless, Batch, and Dedicated Inference for Scalable AI Deployment, 200+ AI Models for Text, Image, Video, Code, and Audio, GPU Clusters With NVIDIA H100, H200, B200, and GB200 Support, Fine-Tuning With SFT, DPO, LoRA, and Reinforcement Learning API and FlashAttention-4 Kernel Delivering 4x Long-Context Performance Improvement. Together these features cover the core workflows most teams expect from a modern ai business tools, from initial setup through day-to-day production use.

Together AI runs as a self-contained product, so you can adopt it without touching the rest of your stack — useful when you want to evaluate the tool in isolation before wiring up integrations.

Who is Together AI for?

Together AI is most useful for AI Engineers, ML Researchers, Backend Developers and AI Startup Teams. If your team falls into one of those buckets, the feature set lines up well with how you already work — you won't be forcing a square peg into a round hole.

Beyond the obvious use case, the product tends to attract users who want a pay-as-you-grow option in the ai business tools space.

Together AI pricing explained

Together AI is a paid product without a permanent free tier. Pricing is positioned around Pay-as-you-go. Most teams justify the cost on time saved or output quality rather than treating it as a discretionary line item.

Across the AI Cloudbase rubric, we score paid pricing models on transparency, rate-limit honesty, and how predictable spend is at scale. Together AI's paid approach is common for production-grade tools and tends to be predictable once you size your usage correctly.

Our verdict on Together AI

Together AI hasn't been rated by enough reviewers yet to publish an aggregate score. The strongest signal in those reviews is that research-backed inference optimization. The most common complaint is that requires developer technical expertise — worth knowing before you commit, but rarely a deal-breaker for teams that already match the use case.

If you're evaluating Together AI against alternatives, weigh it on the same 7-criteria rubric we apply to every tool: capability, integrations, pricing transparency, support, security posture, roadmap velocity, and community signal. Built by Together Computer, Inc., founded in 2022, the product has a clear track record you can verify before adopting it. The bottom line: Together AI is a solid pick in the ai business tools category, and it deserves a spot on your shortlist if your workflow matches what it was built for.

Trusted Reviews

Verified Platforms

4.5 G2 (30) 2.4 Trustpilot (6)

What's New

weekly

Model Redirects For GLM-4.6 And Mistral-7B

Automatic model redirect from GLM-4.6 to GLM-4.7 and Mistral-7B-Instruct-v0.3 to Ministral-3-14B-Instruct-2512 as same-lineage upgrades.

May 9

Pricing Update For Gemma 3n E4B

Updated pricing for google/gemma-3n-E4B-it model: input price changed from $0.02 to $0.06, output price changed from $0.04 to $0.12 per 1M tokens.

Apr 15

View all updates

User Base

500K+ developers

Active Users

Security & Privacy

SOC 2 Type II HIPAA Compliant CCPA Compliant GDPR Compliant

End-to-end encryption (in transit and at rest) Multi-factor authentication (MFA) Role-based access control (RBAC) Zero Data Retention (ZDR) mode Network segmentation and automated threat detection